Cynthia BOUGEARD Haloacetic Acids and other Disinfection By

Transcription

CRANFIELD UNIVERSITY
Cynthia BOUGEARD
Haloacetic Acids and other Disinfection By-Products in
UK Treated Waters: Occurrence, Formation and Precursor
Investigation
School of Applied Sciences
Centre for Water Science
PhD
CRANFIELD UNIVERSITY
School of Applied Sciences
Centre for Water Science
Ph.D. Thesis
Academic Year 2008 – 2009
Cynthia BOUGEARD
Haloacetic Acids and other Disinfection By-Products in UK
Treated Waters: Occurrence, Formation and Precursor
Investigation
Supervisors: Prof. Simon A. PARSONS and Dr. Bruce JEFFERSON
June 2009
This thesis is submitted in partial fulfilment of the requirements for the degree of
Doctor of Philosophy
© Cranfield University, 2009. All rights reserved. No part of this publication may be
reproduced without written permission of the copyright owner.
ABSTRACT
Disinfection by-products (DBPs) in drinking water are formed when natural organic
matter (NOM) that remains after initial treatment reacts with disinfectants, such as
chlorine or chloramines. DBPs, which are of health concern, can take the form of
trihalomethanes (THMs), haloacetic acids (HAAs), haloacetonitriles (HANs),
haloketones (HKs), haloacetaldehydes (HAs), halonitromethanes (HNMs) and a host of
other halogenated DBPs. So far, regulations in the United Kingdom (UK) only
encompass the group of THMs allowing a maximum level of 100 µg/L. HAAs, the
second most prevalent class of DBPs, are currently under consideration by the European
Union to be regulated at 80 µg/L.
Reliable and reproducible quantification methods are required for DBP detection. To
address this need, the presented work includes a comparative study between analytical
devices, which concludes that GC/ECD is the only approach with suitable detection
limits. This work reports an investigation of the DBP formation potential (FP) of waters
from 11 water treatment works (WTWs) at different locations in the UK. Several of
these waters have shown to form significant levels of HAAs and THMs. Furthermore,
other DBPs, such as iodo-THMs (i-THMs), HANs, HKs, HAs and HNMs were
detected. It has also been confirmed that improving the control of these DBPs can be
achieved by using monochloramine instead of free chlorine. A statistical analysis
revealed that THMs correlated well with the HAAs, and as a result the regulatory limit
of 100 µg/L for the THM4 would fail a regulation of 80 µg/L for the nine HAAs.
A number of parameters have been identified, which have particular relevance when
considering the formation of HAAs and THMs in treated waters. Threshold bromide
level was determined beyond which speciation of DBPs shift toward brominated
species. The pH, which significantly affected THMs, was less strongly linked to the
HAAs. The temperature had a consistent impact with a decreasing DBP formation at
lower temperatures. Increasing the contact time with the disinfectant resulted in parallel
first order reaction kinetics of the HAAs and THMs. Finally, the precursors involved in
the formation of DBPs were found to be specific to water sources.
-i-
-ii-
ACKNOWLEDGEMENTS
The work contained in this thesis was made possible through sponsorship by Anglian
Water, Northumbrian Water Limited, Severn Trent Water, United Utilities and
Yorkshire Water.
I would like to thank Prof. Simon A. Parsons for his help and advice throughout my
PhD. Special thanks go to Dr. Emma H. Goslan for her precious help, technical advice,
support and proof-reading most of this thesis. I also would like to thank Dr. Bruce
Jefferson for his advice and help.
I am very grateful to Dr. Howard Weinberg, who has given me endless support, advice,
help and invited me to his university in order to help in the development of this PhD as
well as the development of my personal career.
I would like to thank everyone in the Centre for Water Science, particularly the fellow
DBP people and the technicians. I am also grateful to Imran Janmohamed for his help.
Thanks also to the friends I met during the PhD, with special thanks to the best
housemate Runi and Melanie, the best officemate John and the best volleyball coach
Nicholas. Special thanks to all the friends from the School of Engineering, and Sarita.
Je tiens à vous remercier, spécialement, Papa et Maman, pour votre soutien durant
toutes ces années d’études, ces trois dernières années de doctorat et particulèrement les
deux derniers mois d’écriture, passés en Bretagne. Merci aussi à Charlie pour les
remises en forme. Merci à Michel, Suzanne, la famille Robert, Bérenge, Virgine et Lulu
pour vos mots d’encouragements. Merci à Béa pour toutes les balades à cheval, qui
m’ont permis de garder le cap.
And finally, thanks to Marco, for being so supportive, helping me to get through the
very difficult periods, proof-reading part of this thesis and keeping me sane!!! Vielen
Dank, daß Du jeden Tag für mich da warst und mir geholfen hast mein Gemüt zu
behalten.
- iii -
-iv-
TABLE OF CONTENTS
Abstract..............................................................................................................................i
Acknowledgements..........................................................................................................iii
List of Figures...................................................................................................................ix
List of Tables..................................................................................................................xiii
Abbreviations, Chemical Formula and Notation.............................................................xv
CHAPTER 1 INTRODUCTION.................................................................................1
1.1 Project background ............................................................................................ 1
1.2 Motivation for work........................................................................................... 2
1.3 Project aims and objectives................................................................................ 3
1.4 Thesis plan and publications.............................................................................. 4
CHAPTER 2 LITERATURE REVIEW.....................................................................9
2.1 Introduction........................................................................................................ 9
2.2 Background ...................................................................................................... 10
2.2.1
Aqueous chlorine chemistry ..................................................................... 10
2.2.2
Chloramine chemistry .............................................................................. 11
2.2.3
NOM......................................................................................................... 13
2.3 HAA and THM occurrence and formation from drinking water..................... 14
2.3.1
Toxicity and regulations ........................................................................... 16
2.3.2
Analytical measurement ........................................................................... 18
2.3.3
Typical levels in drinking water ............................................................... 22
2.3.4
Influence of organic precursors ................................................................ 24
2.3.4.1
Water characteristics: TOC/DOC, UV254 and SUVA254 ................... 24
2.3.4.2
NOM fractionation ............................................................................ 26
2.3.5
Influence of inorganic precursors: bromide.............................................. 31
2.3.6
Influence of water treatment variables ..................................................... 33
2.3.6.1
pH ...................................................................................................... 33
2.3.6.2
Temperature ...................................................................................... 34
2.3.6.3
Alternative disinfectants.................................................................... 34
2.3.6.4
Disinfectant dose ............................................................................... 36
2.3.6.5
Reaction time and kinetics ................................................................ 37
2.3.7
Predictive models for HAAs and THMs .................................................. 37
2.4 Semi-volatile DBP occurrence and formation from drinking water................ 44
2.4.1
Toxicity and regulations ........................................................................... 46
2.4.2
Analytical measurement ........................................................................... 47
2.4.3
Typical levels in drinking water ............................................................... 48
2.4.4
Parameters affecting semi-volatile DBP formation.................................. 49
2.5 Chapter summary............................................................................................. 49
-v-
Table of contents
CHAPTER 3 ANALYTICAL METHODOLOGY FOR HAAs ............................51
3.1 Introduction...................................................................................................... 51
3.2 Method comparison for HAA analysis ............................................................ 52
3.2.1
Materials and Methods: IC ....................................................................... 52
3.2.1.1
Reagents and glassware..................................................................... 52
3.2.1.2
Preparation of standards .................................................................... 53
3.2.1.3
Method description............................................................................ 53
3.2.1.4
Instrumentation conditions................................................................ 54
3.2.2
Materials and Methods: GC...................................................................... 54
3.2.2.1
3.2.2.2
3.2.2.3
3.2.2.4
3.2.3
Results and Discussion ............................................................................. 59
3.2.3.1
IC method.......................................................................................... 59
3.2.3.2
GC methods....................................................................................... 61
3.2.4
Conclusion ................................................................................................ 64
3.3 GC/ECD method validation............................................................................. 64
3.3.1
Materials and Methods ............................................................................. 65
3.3.1.1
3.3.1.2
3.3.1.3
3.3.1.4
3.3.1.5
Method performance ......................................................................... 66
3.3.1.6
Water samples ................................................................................... 67
3.3.2
Results and Discussion: Evaluation of method performance ................... 67
3.3.3
Conclusion ................................................................................................ 72
3.4 Chapter summary............................................................................................. 73
CHAPTER 4 ANALYTICAL METHODOLOGY FOR SEMI-VOLATLE
DBPs................................................................................................................................75
4.1 Introduction...................................................................................................... 75
4.2 Qualitative identification of the semi-volatile DBPs ....................................... 77
4.2.1
4.2.1.1
4.2.1.2
4.2.1.3
Water sample preparation.................................................................. 80
4.2.1.4
4.2.1.5
Instrument conditions........................................................................ 82
4.2.2
4.2.3
Conclusion ................................................................................................ 85
4.3 LLE with GC/ECD method validation ............................................................ 85
4.3.1
4.3.1.1
Reagent and glassware ...................................................................... 85
4.3.1.2
4.3.1.3
4.3.1.4
Instrument conditions........................................................................ 88
4.3.1.5
Method performance ......................................................................... 88
-vi-
Table of contents
4.3.1.6
Water samples ................................................................................... 88
4.3.2
4.3.3
Limitations................................................................................................ 97
4.3.4
Conclusion ................................................................................................ 97
4.4 Chapter summary............................................................................................. 97
CHAPTER 5 COMPARISON OF THE DBP-FP OF UK TREATED WATERS
EXPOSED TO CHLORINE AND MONOCHLORAMINE.....................................99
5.1 Introduction...................................................................................................... 99
5.2 Materials and Methods................................................................................... 101
5.2.1
Water sample collection ......................................................................... 101
5.2.2
Water sample characterisation................................................................ 101
5.2.2.1
pH .................................................................................................... 101
5.2.2.2
Non purgeable organic carbon (NPOC) .......................................... 101
5.2.2.3
UV / SUVA ..................................................................................... 102
5.2.2.4
Bromine and iodine measurement................................................... 102
5.2.3
DBP-FP................................................................................................... 102
5.2.3.1
Chlorine and monochloramine solution .......................................... 102
5.2.3.2
Buffer .............................................................................................. 103
5.2.3.3
Quench solutions ............................................................................. 103
5.2.3.4
Samples exposed to chlorine and monochloramine ........................ 103
5.2.4
Sample analytical methods ..................................................................... 104
5.3 Results and Discussion .................................................................................. 104
5.3.1
Water characterisation ............................................................................ 104
5.3.2
DBP levels from different water sources................................................ 106
5.3.2.1
HAAs............................................................................................... 106
5.3.2.2
THMs and i-THMs.......................................................................... 111
5.3.2.3
HANs............................................................................................... 114
5.3.2.4
HKs, HAs and HNMs...................................................................... 116
5.3.3
Relationships between HAAs, THMs and other DBPs .......................... 119
5.3.4
Relationships between NPOC, UV and SUVA with DBPs.................... 121
5.4 Chapter summary........................................................................................... 124
CHAPTER 6 UNDERSTANDING AND CONTROL OF HAA AND THM
FORMATION IN TREATED WATERS..................................................................127
6.1 Introduction.................................................................................................... 127
6.2 Materials and Methods................................................................................... 129
6.2.1
Water sample and characterisation ......................................................... 129
6.2.1.1
Measurement of bromide using IC.................................................. 130
6.2.1.2
Alkalinity......................................................................................... 131
6.2.2
Fractionation........................................................................................... 131
6.2.3
DBP-FP................................................................................................... 133
6.2.3.1
Sample preparation for the determination of parameter impacts and
the modelling study .......................................................................................... 133
6.2.3.2
Quality control samples................................................................... 134
6.2.3.3
Sample preparation for the determination of HAA precursors ....... 134
6.2.3.4
HAA and THM sample preparation ................................................ 134
6.2.4
Sample analytical methods ..................................................................... 135
- vii -
Table of contents
6.3 Results and Discussion .................................................................................. 136
6.3.1
Water characterisation ............................................................................ 136
6.3.2
Parameters affecting the formation of HAAs and THMs....................... 137
6.3.2.1
Impact of pH.................................................................................... 137
6.3.2.2
Impact of bromide ........................................................................... 140
6.3.2.3
Impact of temperature ..................................................................... 142
6.3.2.4
Impact of reaction time ................................................................... 143
6.3.2.5
Impact of disinfectant...................................................................... 144
6.3.3
HAA precursor investigation.................................................................. 147
6.3.4
Modelling of the chlorine decay and prediction of HAA5 and THM4.... 151
6.3.4.1
Chlorine decay data modelling........................................................ 151
6.3.4.2
Chlorine decay to predict THM4 and HAA5 ................................... 154
6.4 Chapter summary........................................................................................... 156
CHAPTER 7 RECOMMENDATIONS FOR WATER UTILITIES ..................159
CHAPTER 8 CONCLUSIONS...............................................................................163
CHAPTER 9 FURTHER WORK...........................................................................167
CHAPTER 10
-viii-
REFERENCES..............................................................................169
LIST OF FIGURES
CHAPTER 1
Figure 1.1
Seasonal variations in HAA formation
3
Figure 1.2
Flow chart of the Phd
7
CHAPTER 2
Figure 2.1
Distribution of chloramine species as a function of pH (Kirmeyer
et al., 2004)
12
Figure 2.2
Breakpoint chlorination scheme (1.0 mg/L nitrogen, temperature
25°C, contact time 2 hours) (Wolfe et al., 1984)
13
Figure 2.3
Correlation between HAAs and THMs
24
Figure 2.4
Relationship between the organic matter concentration
(TOC/DOC) and UV254 for a range of 43 drinking and treated
waters and 32 raw waters
25
Figure 2.5
Schematic of monochloramine reaction pathways in the presence
of NOM (as DOC). Monochloramine reacts directly with reactive
site concentration [DOC1], and free chlorine with reactive site
concentration [DOC2]. Adapted from Duirk and Valentine (2006)
35
Figure 2.6
Observed HAAs against predicted HAAs using the models of Amy
et al. (1998) and Sohn et al. (2004) in raw (A) and treated waters
(B)
43
Figure 2.7
Cytotoxicity and genetoxicity indices for different classes of DBPs
and for chloro-, bromo- and iodo-DBPs (Plewa et al., 2008)
46
Figure 3.1
Sample preparation procedure (Tung et al., 2006) – Adapted from
US EPA Method 552.2 (1995)
56
Figure 3.2
Simplified diagram of GCxGC/TOFMS Instrument
58
Figure 3.3
Chromatogram of HAA9 standards (10 µg/L each). Retention time
(minutes): MCAA = 5.547, MBAA = 5.960, DCAA = 9.767,
BCAA = 11.367, DBAA = 12.103, TCAA = 28.627, BDCAA =
33.590, DBCAA = 40.500, TBAA = 49.873
60
CHAPTER 3
- ix -
List of Figures
Figure 3.4
Comparison of a 100 µg/L of each derived HAA standard
chromatograms from two similar GC/MS instruments by (A) Xie
(2001) and (B) this present study
62
Figure 3.5
(A) Total ion chromatogram for derived HAA6 (equivalent to 100
µg/l of each HAA) and (B) a partially reconstructed mass
chromatogram (m/z = 59) of a derived HAA6 standard at 100 µg/l
of each HAA (Leco Pegasus 4D GC×GC/TOFMS)
63
Figure 3.6
Chromatogram for a diluted (equivalent to 100 µg/l of each) and
derivatised HAA6 standard with GC-ECD
64
Figure 3.7
GC/ECD chromatogram for a derived HAA9 standard (100 µg/L)
(see Table 3.3 for analyte identification)
69
Figure 3.8
Calibration curves for nine HAAs – Concentration: 0.5, 5, 25, 50,
75 and 100 µg/L
70
Figure 4.1
Summary of the SPE – GC/ECD method used for concentrating 16
DBPs in treated water (adapted from Weinberg et al., 2002 and
Chinn et al., 2007)
81
Figure 4.2
Order of elution of THMs (A) and 551B Halogenated Volatiles
Mix (B) using the VF-1ms column (standard at 100 µg/L each –
present study)
83
Figure 4.3
DB-1ms column performance using full DBP set from Weinberg
et al. (2002) (16 targeted compounds + compounds not studied
here)
83
Figure 4.4
Elution of BDCM and DCAN with VF-1ms (A) and ZB-1701 (B)
84
Figure 4.5
Summary of the LLE with GC/ECD method (adapted from
Krasner et al., 2001)
87
Figure 4.6
Individual chromatograms for semi-volatile contaminants and
THMs (100 µg/L each DBP)
89
Figure 4.7
GC/ECD chromatogram for an extracted DBP16 standard (100
µg/L) (see Table 4.2 for analyte identification)
90
Figure 4.8
Calibration curves for (A) DCIM, TBM, BCIM and TCM – (B)
DCA, BDCM, TCA, 1,1-DCP, TCNM and DBCM – (C) TCAN,
DCAN, BCAN, 1,1,1-TCP, DBAN and DBNM – Concentration:
0.5, 3, 7, 15, 50 and 100 µg/L
92
Figure 4.9
Contamination and TCM peaks while working on recovery in
lowland water
96
CHAPTER 4
-x-
List of Figures
CHAPTER 5
Figure 5.1
Distribution of HAAs after 24 hours bench scale exposure to
chlorine and monochloramine for 11 treated waters
108
Figure 5.2
BIF in chlorinated and monochloraminated samples versus
bromine concentration
110
Figure 5.3
Distribution of THMs after 24 hours bench scale exposure to
112
Figure 5.4
Distribution of i-THMs after 24 hours bench exposure to chlorine
and monochloramine for 11 treated waters (ND – not detected)
113
Figure 5.5
Distribution of HANs after 24 hours bench scale exposure to
115
Figure 5.6
Distribution of HKs after 24 hours bench scale exposure to
116
Figure 5.7
Distribution of HAs after 24 hours bench scale exposure to
117
Figure 5.8
Distribution of HNMs after 24 hours bench scale exposure to
chlorine and monochloramine for 11 treated waters (ND – not
detected)
118
Figure 5.9
Correlation between THMs with HAAs and the semi-volatile
DBPs (chlorine FP tests)
120
Figure 5.10
Relationship between NPOC and DBPs
122
Figure 5.11
Relationship between UV and DBPs for coagulated waters
123
Figure 6.1
Process schematic of the lowland WTW (a Anthracite, sand and
garnet)
130
Figure 6.2
Process schematic of the upland WTW (a Dissolved air flotation)
130
Figure 6.3
Schematic of the fractionation procedure
132
Figure 6.4
Comparison of pH effect on the formation of measured HAA5 in
the lowland water
138
Figure 6.5
Comparison of pH effect on the formation of measured HAA5 in
the upland water
139
Figure 6.6
Comparison of pH effect on the formation of THM4 in the lowland
water
140
Figure 6.7
Impact of bromide on the formation of measured HAA5 in the
upland water
141
Figure 6.8
Temperature effect on the DBP-FP in lowland and upland water
142
CHAPTER 6
- xi -
List of Figures
Figure 6.9
Impact of chlorine contact time on HAA5 and THM4 formation
from lowland water (LW) and upland water (UW), experimental
conditions: pH = 7°C, temperature = 20°C, NPOC of lowland
water = 4.7 mg/L and upland water = 2.1 mg/L
144
Figure 6.10
Comparison of chlorine versus chloramines in lowland water, at
pH 6 and pH 8; Cl2:N ratio = 3:1 and 7:1, contact time = 168 hours
145
Figure 6.11
Comparison of chlorine versus chloramines in upland water, at pH
6 and pH 8; Cl2:N ratio = 3:1 and 7:1, contact time = 168 hours
146
Figure 6.12
HAA concentrations from lowland water (LW) and upland water
(UW) from five different fractions; experimental conditions: pH =
7, contact time = 72 hours, temperature = 20°C
150
Figure 6.13
Fractions to be targeted for their removal
151
Figure 6.14
Chlorine decay data of (A) lowland water (LW) and (B) upland
water (UW), both at pH 6, 7 and 8 fitted to parallel first order
reaction model
152
Figure 6.15
HAA5 concentration and formation model predictions for (A)
lowland water (LW) and (B) upland water (UW), both at pH 6, 7
and 8
155
Figure 6.16
THM4 concentration and formation model predictions for (A)
lowland water (LW) and (B) upland water (UW), both at pH 6, 7
and 8
155
-xii-
LIST OF TABLES
CHAPTER 2
Table 2.1
Chemical and physical properties of THMs
14
Table 2.2
Chemical and physical properties of HAAs
16
Table 2.3
Analytical method, description, and limit of detection
20
Table 2.4
Reported levels of HAAs and THMs in drinking and/or treated
water
23
Table 2.5
Guidelines for the nature of NOM (Edwald and Tobiason, 1999)
24
Table 2.6
Relationship between HAA-FP
characteristics in treated waters
water
26
Table 2.7
Fractions, their characteristics, their chlorine demand and their
DBP-FP
29
Table 2.8
Summary of NOM properties (Hwang et al., 2001)
31
Table 2.9
Effect of bromide on the molar yield of THMs and HAAs
(reaction conditions: pH = 7, contact time = 48 h, temperature =
20°C, Cl2 dose = 6.2 mg/L and initial water source bromide
concentration = 63 µg/L) (Hua et al., 2006)
32
Table 2.10
Selected HAA predictive models
39
Table 2.11
Chemical and physical properties of selected semi-volatile DBPs
45
Table 2.12
Typical levels of semi-volatile DBPs found in drinking water
48
Table 3.1
Calibration correlation coefficients for HAA9 analysed by IC in
this present study and in a study by Liu and Mou (2003)
59
Table 3.2
Detection limits for the analytes from this present study, Liu and
Mou (2003) and Nair et al. (1994)
61
Table 3.3
Identification of the compounds and method performance
67
Table 3.4
Spike recovery in lowland and upland water fortified with HAAs
72
Table 3.5
Summary of
performance
73
and
THM-FP
and
CHAPTER 3
analytical
method
conditions
and
methods
- xiii -
List of Tables
CHAPTER 4
Table 4.1
Compound purity and suppliers
78
Table 4.2
Identification of the compounds and method performance
91
Table 4.3
Spike recovery in lowland water fortified with DBPs
94
Table 4.4
Spike recovery in upland water fortified with DBPs
95
CHAPTER 5
Table 5.1
List of WTWs, sources and water characteristics
105
Table 5.2
Correlation between DBPs and water characteristics
123
Table 5.3
Summary of the findings
125
Table 6.1
Water characteristics
136
Table 6.2
Decrease of THM4 in chloraminated lowland and upland water
compared to the same chlorinated sample; contact time with
chloramines = 168 hours
147
Table 6.3
Summary of NOM fraction properties in lowland and upland water
148
Table 6.4
Lowland and upland water chlorine decay constants for parallel
first order reaction model
153
Table 6.5
THM4 and HAA5 yield coefficients at three pHs and model
coefficient of correlation R2
156
CHAPTER 6
-xiv-
ABBREVIATIONS, CHEMICAL FORMULA AND NOTATION
Abbreviations and Chemical Formula
1,1-DCP
1,1- dichloropropanone
1,1,1-TCP
1,1,1- trichloropropanone
amu
Atomic Mass Unit
APHA
American Public Health Association
ASG
Anthracite, Sand and Garnet
AWWA
American Water Works Association
AwwaRF
Awwa Research Foundation
BCAA
Bromochloroacetic Acid
(µg/l)
BCAN
Bromochloroacetonitrile
(µg/l)
BCIM
Bromochloroiodomethane
(µg/l)
BDCAA
Bromodichloroacetic Acid
(µg/l)
BDCM
Bromodichloromethane
(µg/l)
BDIM
Bromodiiodomethane
(µg/l)
BIF
Bromine Incorporation Factor
-
Br
Bromide Ion
CDIM
Chlorodiiodomethane
CE
Capillary Electrophoresis
CH3CN
Acetonitrile
CHO
Chinese Hamster Ovary
Cl-
Chloride Ion
Cl2
Chlorine
CO2
Carbon Dioxide
CZE
Capillary Zone Electrophoresis
D/DBPR
Disinfectants/Disinfection by-Product Rule
DAF
Dissolved Air Flotation
(µg/l)
- xv -
Abbreviations, Chemical Formula and Notation
DBAA
Dibromoacetic Acid
(µg/l)
DBAN
Dibromoacetonitrile
(µg/l)
DBCAA
Dibromochloroacetic Acid
(µg/l)
DBCM
Dibromochloromethane
(µg/l)
DBIM
Dibromoiodomethane
(µg/l)
DBNM
Dibromonitromethane
(µg/l)
DBP
Disinfection By-Product
(µg/l)
DBP-FP
Disinfection By-Product Formation Potential
DCA
Dichloroacetaldehyde
(µg/l)
DCAA
Dichloroacetic Acid
(µg/l)
DCAN
Dichloroacetonitrile
(µg/l)
DCBM
Dichlorobromomethane
(µg/l)
DCIM
Dichloroiodomethane
(µg/l)
DNA
Deoxyribonucleic Acid
DOC
Dissolved Organic Carbon
(mg/L)
DON
Dissolved Organic Nitrogen
(mg/L)
DVB
Divinyl Benzene
DWI
Drinking Water Inspectorate
DXAA
Dihalogenated Acetic Acid
ED
Electrochemical Detector
EI
Electron Ionisation
EOM
Extracellular Organic Matter
EPA
Environmental Protection Agency
ESI/MS
Electrospray Ionization/ Mass Spectrometry
FP
Formation Potential
GAC
Granular Activated Carbon
GC
Gas Chromatography
GC/ECD
Gas Chromatography/Electron Capture Detector
GC/MS
Gas Chromatography/Mass Spectrometry
GC/TOFMS
Gas Chromatography/Time of Flight Mass Spectrometry
-xvi-
(µg/l)
H+
Hydrogen Ion
H2CO3
Carbonic Acid
H2SO4
Sulphuric Acid
H3O+
Hydronium Ion
HA
Haloaldehyde
(µg/l)
HAA
Haloacetic Acid
(µg/l)
HAA3
Sum of three HAAs
(µg/l)
HAA5
Sum of five HAAs
(µg/l)
HAA6
Sum of six HAAs
(µg/l)
HAA9
Sum of nine HAAs
(µg/l)
HAA-FP
Haloacetic Acid Formation Potential
HAN
Haloacetonitrile
HCl
Hydrogen Chloride
HK
Haloketone
-
Hypobromous Acid/Hypobromite Ion
-
Hypochlorous Acid/Hypochlorite Ion
HOBr/OBr
HOCl/OCl
HOI
Hypoiodous Acid
HPI-A + N
Hydrophilic Acid + Neutral
HPI-B
Hydrophilic Base
HPI-N
Hydrophilic Neutral
HPLC
High Performance Liquid Chromatography
HPO-A
Hydrophobic Acid
HPO-B
Hydrophobic Base
HPO-N
Hydrophobic Neutral
HPSEC
High Performance Size Exclusion Chromatography
Hz
Hertz
IC
Ion Chromatography
ICP/MS
IO3
-
(µg/l)
(µg/l)
Inductively Coupled Plasma/Mass Spectrometry
Iodate
IS
Internal Standard
i-THM
Iodo-THM
- xvii -
KH2PO4
Potassium Acid Phosphate
LLE
Liquid-Liquid Extraction
LOD
Limit of Detection
m/z
Mass to charge ratio
MBAA
Monobromoacetic Acid
(µg/l)
MCAA
Monochloroacetic Acid
(µg/l)
MCL
Maximum Contaminant Level
MRL
Maximum Reporting Level
MtBE
Methyl tert Butyl Ether
MW
Molecular Weight
MX
3-chloro-4-(dichloromethyl)-5-hydroxy-2-(5H)-furanone
MXAA
Monohalogenated Acetic Acid
NaHCO3
Sodium Bicarbonate
NaOCl
Sodium Hypochlorite
Na2CO3
Sodium Carbonate
Na2HPO4
Phosphate Dibasic
Na2SO4
Sodium Sulphate
NCI
National Cancer Institute
NCl3
Trichloramine
NDMA
N-nitrosodimethylamine
NHCl2
Dichloramine
NH2Cl
Monochloramine
NH3
Ammonia
NH4Cl
Ammonium Chloride
NH4OH
Ammonium Hydroxyde
NOM
Natural Organic Matter
NPOC
Non Purgeable Organic Carbon
pH
potential Hydrogen
PHB
Polyhydroxybutyrate
-xviii-
(µg/l)
(mg/L)
P&T
Purge and Trap
PTFE
Polytetrafluoroethylene
RSD
Relative Standard Deviation
SPE
Solid Phase Extraction
SUVA
Specific Ultraviolet Absorbance
(L/m-mg C)
SUVA254
Specific Ultraviolet absorbance at 254 nm
(L/m-mg C)
SV
Semi-Volatile
TBAA
Tribromoacetic Acid
(µg/l)
TBM
Tribromomethane
(µg/l)
TC
Total Carbon
TCA
Trichloroacetaldehyde
(µg/l)
TCAA
Trichloroacetic Acid
(µg/l)
TCAN
Trichloroacetonitrile
(µg/l)
TCM
Trichloromethane
(µg/l)
TCNM
Trichloronitromethane
(µg/l)
THM
Trihalomethane
(µg/l)
THM4
Sum of four THMs (TCM, BDCM, DBCM and TBM)
THM-FP
Trihalomethane Formation Potential
TIM
Triiodomethane
(µg/l)
TOBr
Total Organic Bromine
(µg/l)
TOC
Total Organic Carbon
(mg/L)
TOCl
Total Organic Chlorine
(µg/l)
TOI
Total Organic Iodine
(µg/l)
TOX
Total Organic Halogen
(µg/l)
TPI
Transphilic
TXAA
Trihalogenated Acetic Acid
UK
United Kingdom
UNC
University of North Carolina
US
United States
US EPA
United States Environmental Protection Agency
(µg/l)
- xix -
UV
Ultraviolet
(1/m)
UV254
Ultraviolet absorbance at 254 nm
(1/m)
WHO
World Health Organisation
WTW
Water Treatment Works
Notation
A
Reactant for kinetic reaction
α
THM yield coefficient
ADBP
Peak area of DBP
AHAA
Peak area of HAA
AIS
Peak area of internal standard
At
Millilitre standard acid titrant
B
Reactant for kinetic reaction
β
HAA yield coefficient
C0
Initial chlorine concentration
(mg/L)
C(t)
Chlorine concentration at any time
(mg/L)
CDBP
Concentration of DBP
(µg/l)
CHAA
Concentration of HAA
(µg/l)
D
Concentration fortified sample
(µg/l)
E
Concentration unfortified sample
(µg/l)
F
Fortified concentration
(µg/l)
f
Fraction of chlorine demand attributed to rapid reactions
k1
Constant of 1st order kinetic reaction
k2
Constant of 2nd order kinetic reaction
k’
Column capacity factor
kR
1st order rate constant for rapid reactions
(h-1)
kS
1st order rate constant for slow reactions
(h-1)
N
Normality
n
Number of replicates
R2
Coefficient of correlation
-xx-
(µg THM/mg Cl2)
(µg HAA/mg Cl2)
S
Standard deviation of replicate analysis
T
Temperature
t
Time
t(n-1,1-α=0.99)
Student t value for the 99% interval confidence level with n-1 degrees
of freedom
(ºC)
(hours)
- xxi -
-xxii-
Chapter 1
INTRODUCTION
1.1
PROJECT BACKGROUND
Disinfection by-products (DBPs) were first discovered in the 1970s in chlorinated
drinking water (Bellar et al., 1974; Rook, 1974) and consequently, the use of chlorine as
a disinfectant has come under investigation. Chlorine reacts with the natural organic
matter (NOM) ubiquitous in water to form DBPs, which can take the form of
trihalomethanes (THMs), haloacetic acids (HAAs), haloacetonitriles (HANs),
haloketones (HKs), haloacetaldehydes (HAs), halonitromethanes (HNMs) and a host of
other halogenated DBPs (Richardson et al., 2007). Many of these DBPs have been
shown to be cytotoxic, genetoxic or to cause cancer in laboratory animals, and are as
such a public health concern (Bull et al., 1990; Plewa et al., 2002).
Richardson (1998) reported approximately 500 DBPs in drinking water of which the
most prevalent groups are the THMs and the HAAs (Singer, 2002). In the United
Kingdom (UK), THMs are the only regulated DBPs and it is required by law that the
sum of four THMs (trichloromethane – TCM, bromodichloromethane – BDCM,
dibromochloromethane – DBCM and tribromomethane – TBM) does not exceed 100
µg/L with a frequency of sampling dependent on the population size (DWI, 2005). In
the United States (US), the US Environmental Protection Agency (US EPA) has
established maximum contaminant levels (MCLs) of 80 µg/L for four THMs and 60
µg/L for five HAAs (monochloro-, monobromo-, dichloro-, trichloro- and
dibromoacetic acid (MCAA, MBAA, DCAA, TCAA and DBAA)) (US EPA, 1998).
HAAs are currently not regulated in the UK, however, the European Union is
considering regulating the nine HAAs at 80 µg/L (Cortvriend, 2008) and as such there is
growing interest in the levels of these compounds found in UK drinking waters and how
best to control them.
One option which would help the water utilities to comply with these proposed
regulations, would be to use monochloramine as a secondary disinfectant, as it
1
Chapter 1 – Introduction
has been shown to reduce DBP formation and also, like chlorine provides a residual in
water distribution systems. However, the use of monochloramine may also lead to an
increase in other DBPs such as HANs and i-THMs (Krasner et al., 1989; Bichsel and
Von Gunten, 2000). HANs and i-THMs are two unregulated classes of semi-volatile
DBPs also present in disinfected waters alongside other unregulated DBPs including
HNMs, HAs and HKs (Krasner et al., 2006). Many of these DBPs have been reported to
be more cytotoxic and genetoxic than some of the regulated DBPs, and the brominated
DBPs are in general more genetoxic and carcinogenic than are the chlorinated
compounds, and the iodinated DBPs are the most genotoxic (Richardson et al., 2007;
Plewa et al., 2008). Little is known about the formation of semi-volatile DBPs in UK
waters.
1.2
MOTIVATION FOR WORK
The work presented in this thesis brought together five UK water companies, concerned
by emerging DBPs: Anglian Water, Northumbrian Water Limited, Severn Trent Water,
United Utilities and Yorkshire Water.
In 2005, a study on HAAs and THMs has been commissioned by the five water
companies cited above plus Thames Water, with objective to determine the relative
distribution and speciation of THMs and HAAs in a variety of UK water treatment
works (WTWs) (confidential, not published). Sampling was undertaken seasonally, and
both groups of species were measured in WTW final treated water and distribution
samples. In addition, THM and HAA formation potential (FP) was determined under
controlled uniform conditions. Initial results showed that some final waters had
concentrations of HAAs between 180 and 265 µg/L (Figure 1.1), well above the
regulations promulgated in the US, and as such identified a need for a more detailed
investigation.
2
HAA (µg/L)
1000
100
Autumn
Winter
Spring
Summer
10
1
0
5
10
15
20
25
WTW
Figure 1.1 Seasonal variations in HAA formation
The main conclusion from this work was that many water quality and treatment factors
influence the distribution of THMs and HAAs in drinking water. Nevertheless, it was
difficult to determine any strong relationship between any water parameters and HAA
and THM formation. Another UK study on HAAs and THMs, undertaken by Malliarou
et al. (2005), reported HAAs with mean levels in the UK ranging from 35 to 95 µg/L
and a maximum concentration of 244 µg/L.
To investigate further the formation and occurrence of DBPs, it was deemed necessary
to set up convenient and sensitive methods for the determination of DBPs in UK
drinking water. Clearly, effective monitoring of DBPs should enable a better
understanding of DBPs and hence, lead to an optimisation of water treatments and
better control of these DBPs.
1.3
PROJECT AIMS AND OBJECTIVES
The purpose of this thesis was to broaden the understanding of DBP-FP of UK treated
waters with reliable and reproducible methods in order to optimise their control in
WTWs and therefore, to meet THM regulations and to anticipate the HAA ones.
3
Consequently, a series of objectives were identified:
 To determine the best analytical method for the measurement of HAAs in UK
treated waters.
 To broaden the number of DBPs analysed and therefore, set up reliable and
reproducible method for their determination.
 To compare the impact of the choice of disinfectant on the level of DBP formed.
 To assess the important parameters affecting the formation of THMs and HAAs.
 To determine if correlations exist between water characteristics and DBP-FP.
 To determine precursors for HAAs and possible common precursors between
waters.
1.4
THESIS PLAN AND PUBLICATIONS
Initially, a literature review was carried out (Chapter 2) to determine methods of
analysis to determine HAAs and other DBPs in treated and drinking waters. This
literature review also emphasises the formation, occurrence and precursors of DBPs in
the US waters, but shows a lack of data for the UK waters, where water treatments are
not necessarily similar to those applied in US WTWs.
Chapter 3 to Chapter 6 cover the technical content of this thesis and the link between
the chapters are presented in a flow chart (Figure 1.2). Chapter 3 assess the different
analytical methods available for the measurement of HAAs. Four methods have been
investigated, from which the most reliable and reproducible has been chosen and
validated for quantification of HAAs in UK treated waters. The first part of this chapter
describing the method comparison has been presented as a paper at an international
conference (Bougeard, C. M. M., Janmohamed, I. H. S., Goslan, E. H., Jefferson, B.,
Watson, J. S., Morgan, G. H. and Parsons, S. A. (2007). Occurrence of Trihalomethanes
(THMs) and Haloacetic Acids (HAAs) in UK waters. In: 233rd National Meeting of
American Chemical Society Division of Environmental Chemistry, 27-29 March,
Chicago, IL, US.) and later published: Bougeard, C. M. M., Janmohamed, I. H. S.,
Goslan, E. H., Jefferson, B., Watson, J. S., Morgan, G. H. and Parsons, S. A. (2008). In:
4
Disinfection By-Products in drinking Water: Occurrence, Formation, Health Effects,
and Control. American Chemical Society Symposium Series 995, edited by Karanfil,
T., Krasner, S. W., Westerhoff, P. and Xie, Y. Washington, DC, US, p. 95-108. The
first part of the book chapter highlighted which methods fitted the best to the
quantification of HAAs in treated waters.
Chapter 4 shows the interest of the author in DBPs other than HAAs, and demonstrates
that i-THMs, HANs, HAs, HKs and HNMs could also be found in UK treated waters.
Therefore, in this chapter, it is shown the qualitative identification of these compounds
and the development of an analytical method for their quantification. This work was
presented at an international conference as a paper (Bougeard, C. M. M., Goslan, E. H.,
Jefferson, B., Parsons, S. A. (2008). Disinfection by-products (DBPs) in UK drinking
water. In: International Conference on Emerging Issues in Disinfection By-Products, 3rd
April, Cranfield, UK). The method developed here was also published jointly with
Chapter 5.
Chapter 5 reports an investigation of 11 WTWs across England and Wales for their
potential to form HAAs, THMs and other halogenated DBPs using chlorine and
preformed monochloramine. To allow this work, methods developed and set up in
Chapter 3 and Chapter 4 were used. This chapter has been submitted as paper format,
jointly with the method description of Chapter 4: Bougeard, C. M. M, Goslan, E. H.,
Jefferson, B., Parsons, S.A. Comparison of the disinfection by-product formation
potential of treated waters exposed to chlorine and monochloramine. Water Research.
Chapter 6 is a more detailed study on the formation of HAAs and THMs. Because
THMs are regulated and HAAs are under consideration for regulation, it was important
to determine the parameters that affect the most their formation. This work was
presented alongside the results from Chapter 3 at the 233rd National Meeting of
American Chemical Society Division of Environmental Chemistry, and later published
in the same book chapter cited above. An investigation of the precursor reactivity
towards the formation of HAAs is also investigated in this chapter. Finally, a
mathematical approach is presented to study the kinetics of chlorine decay and a model
was adapted to UK treated waters for the prediction of THMs and HAAs. The work
related to the chlorine decay was part of a poster presentation at an international
5
conference (Bougeard, C. M. M, Goslan, E. H., Jefferson, B., Parsons, S.A. (2006). In:
Gordon Conference in Drinking Water Disinfection By-Products, 13-18 August, South
Hadley, MA, US).
The overall implications of this research are pulled together in Chapter 7
“Recommendations for water utilities” and key conclusions from this thesis are made
(Chapter 8).
Suggestions for further work are reported in Chapter 9.
6
Past research on UK DBPs
Cranfield Survey (2005,
confidential)
Malliarou et al., (2005)
Further work needed on
HAAs for better
understanding and better
control
Set-up analytical method for
HAA quantification in UK
treated waters
CHAPTER 3
Growing interest in the
scientific community for the
semi-volatile DBPs, because
of potential toxicity
Identification semi-volatile DBPs
Set-up analytical method for their
quantification in UK treated waters
CHAPTER 4
Quantification of HAAs, THM and
semi-volatile DBPs in 11 UK treated
waters; disinfectant comparison
CHAPTER 5
THMs regulated at 100 µg/L
HAAs under considerations
Detailed investigation of HAAs and
THM in UK treated waters to allow
better control in water utilities
CHAPTER 6
Recommendations for water utilities
CHAPTER 7
Conclusions
CHAPTER 8
Further research
CHAPTER 9
Figure 1.2 Flow chart of the Phd
7
8
Chapter 2
LITERATURE REVIEW
2.1
INTRODUCTION
DBPs in drinking water were first discovered in the 1970s and the first public health
concern was raised with the identification of TCM and other THMs in chlorinated water
(Bellar et al., 1974; Rook, 1974). Since that, more than hundreds of DBPs have been
found in drinking water (Richardson et al., 1998). The most prevalent chlorinated DBPs
are the THMs and the HAAs, comprising more than 50% on a weight basis, followed by
trichloroacetaldehyde (TCA - HAs), HANs, HKs and HNMs (Singer et al., 2002). All
these chemicals, plus the iodinated counterparts of THMs (i-THMs), have been detected
in drinking water that has been treated using the most common disinfectants (chlorine,
chloramines, ozone and chlorine dioxide) (Krasner et al., 2006).
Because of their potential health concern, HAAs and THMs are regulated in the US at
60 µg/L and 80 µg/L respectively (US EPA, 1998). In the UK, only THMs are regulated
at a concentration of 100 µg/L (DWI, 2005). To date, no regulations have been
promulgated for HAs, HANs, HKs, HNMs and i-THMs, despite their toxicity (Plewa et
al., 2004; Plewa et al., 2008), but the World Health Organisation (WHO) has suggested
guideline values of 20 µg/L for dichloroacetonitrile (DCAN), 70 µg/L for
dibromoacetonitrile (DBAN) and 10 µg/L for TCA (WHO, 2006).
DBPs are formed when NOM that remains after treatment processes reacts with the
disinfectants. Therefore, to have a better understanding of DBP formation and
occurrence, a significant amount of research has been conducted with the aim to identify
the DBP levels and their precursors, the reaction pathways and the physical parameters
that lead to DBP formation. DBP precursors can be organic, such as the nature and
concentration of NOM, or inorganic, such as bromide and/or iodide present in the
natural water source (Cowman and Singer, 1996; Bichsel and Von Gunten, 2000;
Hwang et al., 2001; Goslan et al., 2002). Physical parameters, such as the pH, the type
of disinfectants, the contact time, the temperature and the disinfectant dose have also
9
Chapter 2 – Literature Review
been extensively studied and are well known to alter the formation of HAAs and THMs
(Carlson and Hardy, 1998; Singer, 1999; Diehl et al., 2000; Malliarou et al., 2005).
The aim of this review is to give insight into the occurrence and the formation of HAAs,
THMs and other DBPs, present in drinking and/or treated water, with a particular
emphasis on the precursors and the factors involved in the formation of HAAs and
THMs. Analytical methods for the determination of DBPs in drinking and/or treated
water are also assessed. Hence, by determining suitable analytical methods and the
parameters that affect the most the DBP formation in drinking and/or treated water, it is
hope to aid the water companies to adapt their treatments to minimise or to better
control the DBP formation.
2.2
BACKGROUND
2.2.1
Aqueous chlorine chemistry
Chlorine (from the Greek word “χλωρóς” (khlôros), meaning pale green), under
standard pressure and temperature, is a pale green gas (Cl2 or dichlorine), characterised
by a disagreeable suffocating odor that is poisonous (WHO, 1998). Chlorine exists also
predominantly in nature as chloride ion (Cl-), a trace component of all the earth’s
geological compartment other than the oceans, its primary sink (Winterton, 2000).
Chlorine, in the form of gaseous chlorine (Cl2) or sodium hypochlorite (NaOCl), is a
powerful oxidant and is used as water purification, despite the formation of potentially
harmful DBPs associated with its use (Deborde and Von Gunten, 2008).When chlorine
gas is dissolved in water, it hydrolyses rapidly to yield hypochlorous acid (HOCl) as:
Cl 2  H 2 O  HOCl  H   Cl  .
Equation 2.1
Hypochlorous acid is also formed when sodium hypochlorite (NaOCl) is used as the
source of chlorine:
NaOCl  H 2 O  HOCl  Na   OH  .
Equation 2.2
Hypochlorous acid is a weak acid (pKa = 7.6) and dissociates partially into hydrogen
ion (H+) and hypochlorite ion (OCl-) in the reversible reaction:
HOCl  H   OCl  .
10
Equation 2.3
Under typical water treatment conditions in the pH range 6 to 9, hypochlorous acid and
hypochlorite ion, both will be present in the water in a proportion dependent of the
temperature (Deborde and Von Gunten, 2008). Hypochlorous acid is a more effective
disinfectant than hypochlorite ion. Another reaction that occurs in waters containing
bromide ion (Br-) and hypochlorous acid is the formation of hypobromous acid (HOBr)
(Von Gunten and Oliveras, 1998):
HOCl  Br   HBrO  Cl  .
Equation 2.4
Hypobromous acid and hypobromite ion (OBr-) have a pKa of 8.7.
In any case, chlorine or bromine substitution can lead to the formation of DBPs when
NOM is present in the water.
2.2.2
Chloramine chemistry
Chloramine chemistry can be summarised by three reversible reactions involving
ammonia (NH3), hypochlorous acid, monochloramine (NH2Cl), dichloramine (NHCl2),
and trichloramine (NCl3) (Diehl et al., 2000):
NH 3  HOCl  NH 2 Cl  H 2O ,
Equation 2.5
NH 2 Cl  HOCl  NHCl 2  H 2 O ,
Equation 2.6
NHCl 2  HOCl  NCl3  H 2O .
Equation 2.7
The competing reactions are highly dependent upon pH, the ratio of chlorine to
ammonia-nitrogen (or chlorine to nitrogen ratio expressed as Cl2:N), and to a lesser
degree, temperature and contact time. A typical distribution of chloramine species as a
function of pH is shown (Figure 2.1). Experimental conditions were 2.5 mg/L of
chlorine in contact with 0.5 mg/L of ammonia (5:1 Cl2:N weight ratio) for two hours
(Kirmeyer et al., 2004). The authors observed that, for the pH range normally
encountered in drinking water treatment (optimum pH condition 6.5 – 9), almost all of
the chloramine species consist of monochloramine, while at pH less than 6,
dichloramine generally accounts for 20 percents or more of the chloramines formed.
The maximum synthesis of dichloramine occurs at a pH range of 4 to 6 and Cl2:N
weight ratio of 5:1 to 7.6:1.
11
100
NH2Cl
Total combined chlorine (%)
80
60
40
20
NHCl2
NCl3
0
3
4
5
pH
6
7
8
Figure 2.1 Distribution of chloramine species as a function of pH (Kirmeyer et al.,
2004)
The formation of trichloramine (Equation 2.7) occurs predominantly at pH values less
than 4.4 (Figure 2.1) or at Cl2:N weight ratio greater than 7.6:1 (Kirmeyer et al., 2004).
Trichloramine in drinking water systems may be formed at trace levels following the
breakpoint chlorination (White, 1999).
Breakpoint chlorination
The breakpoint chlorination occurs when enough chlorine has been added to meet the
chlorine demand of the water and oxidises all the ammonia. Therefore, chlorine that is
added beyond that point (theoretically a weight ratio of 7.6:1 Cl2:N) remains present in
the water as free chlorine residual and mono-, di- and trichloramine may be only present
as trace levels (Wolfe et al., 1984). Summary of the residual species present in water as
a function of Cl2:N ratio is presented below (Figure 2.2).
12
Residual chlorine as Cl2 (mg/L)
6
NH2Cl residual
present
5
NHCl2 increases
NH2Cl residual
present
4
Free chlorine (HOCl/OCl) are
predominant residual
Trace of NH2Cl, NHCl2 and
NCl3
Cl2:N
5:1
3
Breakpoint
Cl2:N
7.6:1
2
1
0
0
1
2
3
4
5
6
7
8
Chlorine dose (mg/L)
9
10
11
12
13
Figure 2.2 Breakpoint chlorination scheme (1.0 mg/L nitrogen, temperature 25°C,
contact time 2 hours) (Wolfe et al., 1984)
Monochloramine, the only used disinfectant amongst chloramine species, has a much
higher CT value1 than free chlorine and is therefore a poor primary disinfectant (WHO,
2000). However, because of its persistence, it is an attractive secondary disinfectant for
the maintenance of a stable distribution system residual. Di- and trichloramine are too
unstable and highly malodorous to be used as disinfectants.
2.2.3
NOM
NOM, ubiquitous in drinking water, is of concern because it serves as precursor to the
formation of DBPs. It has been extensively studied, but still remains a complex and
heterogeneous mixture of specific but mostly hard-to-identify compounds and varies
significantly from one source to another (Hwang et al., 2001). The same author defined
NOM as a mixture of two separate fractions: the hydrophobic (non-polar) substances,
generally of terrestrial origin and the hydrophilic (polar) substances, typically of
1
The CT value is the product of the disinfectant concentration C in mg/L and the contact time T in
minute required to inactivate a specified percentage of microorganisms (WHO, 2000).
13
biological origin. To better understand the reactivity of NOM towards the formation of
DBPs, NOM is generally characterised by measuring its total organic carbon (TOC) or
dissolved organic carbon (DOC) concentration, its ultraviolet (UV) absorbance,
generally at 254 nm to exhibit the amount of aromatic material, and its potential to form
DBPs (see Section 2.3.4.1). The fractionation and isolation of DBP NOM precursor
with XAD 8 / 4 resins have also been used to determine which fractions of NOM
contribute the most to the DBP formation (Oliver and Thurman, 1983; Reckhow and
Singer, 1990; Croué et al., 1993). However, most of these works focused on NOM from
untreated water and therefore, a lack of results remains on NOM fractions that remain
after treatment in drinking water and their reactivity towards DBPs. An overview on
drinking water NOM fractions and their impact on DBPs are given in Section 2.3.4.2.
2.3
HAA
AND
THM
OCCURRENCE
AND
FORMATION
FROM
DRINKING WATER
The first DBPs to be identified and regulated in drinking water were the THMs
(Symons,
1999).
These
bromodichloromethane
are
trichloromethane
(BDCM),
(TCM
or
dibromochloromethane
chloroform),
(DBCM)
and
tribromomethane (TBM or bromoform). Their molecular formula, their molecular
weight, their structure, their boiling point and their CAS number are given (Table 2.1).
These four THMs are often referred as THM4.
Table 2.1 Chemical and physical properties of THMs
THM
Molecular
formula
Molecular
weight
(g/mol)
Structure
Cl
TCM
CHCl3
CHBrCl2
119.38
CHBr2Cl
14
CHBr3
67-66-3
87
75-27-4
119-120
124-48-1
146-150
75-25-2
Br
163.83
Cl
Br
208.28
Br
Br
TBM
61-62
Cl
Cl
DBCM
CAS number
Cl
Cl
BDCM
Boiling point
(°C)
Br
252.73
Br
The second most abundant DBPs found in drinking water are the HAAs and there are
nine chloro- and bromo-HAAs in total (Singer, 2002). They are classified as
monohalogenated acetic acid (MXAA or XHAA) (monochloroacetic acid (MCAA) and
monobromoacetic acid (MBAA)), dihalogenated acetic acid (DXAA or X2AA)
(dichloroacetic acid (DCAA), dibromoacetic acid (DBAA) and bromochloroacetic acid
(BCAA)) and trihalogenated acetic acid (TXAA or X3AA) (trichloroacetic acid
(TCAA), bromodichloroacetic acid (BDCAA), dibromochloroacetic acid (DBCAA) and
tribromoacetic acid (TBAA)). US regulation is currently in place for five of the HAAs.
These are MCAA, MBAA, DCAA, TCAA and DBAA and are often referred as HAA5.
All nine HAAs are referred as HAA9 for measurement purpose; nevertheless, six HAAs
are often determined (HAA5 + BCAA) and are called HAA6. Their molecular formula,
their molecular weight, their structure, their boiling point, their respective methyl ester
boiling point and their CAS number are given (Table 2.2). The boiling point of methyl
esters is shown as the HAAs are converted to their methyl ester form before gas
chromatography (GC) analysis (Chapter 3).
15
Table 2.2 Chemical and physical properties of HAAs
HAA
Molecular
formula
Molecular
weight
(g/mol)
MCAA
C2H3ClO2
94.45
Boiling
point
(°C)
Boiling
point of
ester (°C)
CAS
number
H
189
130
79-11-8
H
206-208
132
79-08-3
H
194
143
79-43-6
H
215
174
5589-96-8
H
196
168
76-03-9
H
128-130
NRa
631-64-1
H
NRa
NRa
71133-14-7
H
NRa
NRa
5278-95-5
H
245
225
75-96-7
Structure
O
Cl
O
O
MBAA
C2H3BrO2
138.95
Br
O
Cl
DCAA
C2H2Cl2O2
128.94
O
Cl
O
Br
BCAA
C2H2BrClO2
173.39
O
Cl
O
Cl
Cl
TCAA
C2HCl3O2
O
163.39
Cl
O
Br
DBAA
C2H2Br2O2
217.84
O
Br
O
Cl
Cl
BDCAA
C2HBrCl2O2
O
207.84
Br
O
Cl
Br
DBCAA
C2HBr2ClO2
O
252.29
Br
O
Br
Br
TBAA
C2HBr3O2
O
296.74
Br
O
a
Not reported.
2.3.1
Toxicity and regulations
DBPs were first discovered in drinking water in 1974 (Bellar et al., 1974; Rook, 1974)
in the form of THMs. As a response, toxicological evaluations have been undertaken
and first concern for public health arised. Indeed, in 1976, a National Cancer Institute
(NCI) study was released in which TCM (chloroform) was classified as a suspected
human carcinogen. Following this finding, studies on rats and mice linked THMs to
cancers of the colon and the kidneys and HAAs to liver tumours (Dunnick, 1985; Bull
et al., 1990; Boorman et al., 1999). Kargalioglu et al. (2002) also reported MCAA,
DCAA, TCAA, BCAA, DBAA and TBAA to be cytotoxic (toxic to cells) and
16
mutagenic (modify the genetic material) in Salmonella typhimurium. There is a general
opinion that the brominated HAAs are more cytotoxic and genetoxic (cause damage to
deoxyribonucleic acid – DNA) than their chlorinated analogues while working with
Chinese hamster ovary (CHO) cells (Plewa et al., 2002).
The risk of human cancer from DBPs, especially cancers of the colon, the rectum and
the bladder have been documented by several epidemiological studies (King and Marret,
1996; Doyle et al., 1997; Koivusalo et al., 1997; Cantor et al., 1998; Hidelsheim et al.,
1998). There are ongoing concerns that the types of cancer observed in animal studies
for the DBPs that have been tested do not correlate with the types observed in human
epidemiological studies (bladder, colon cancer) (Richardson, 1998). Nevertheless,
several epidemiological studies have reported DBPs to be potentially harmful to
foetuses, such as THMs being linked to the frequency of stillbirths or having a strong
association to spontaneous abortions (Waller et al., 1998; King et al., 2000; Bove et al.,
2002).
To respond to these toxicological and epidemiological studies, the US EPA was actively
prompted to publish regulation and started in 1979 by setting a MCL for THMs as
annual average of 100 µg/L. In 1998, the US EPA strengthened existing rules with its
Stage 1 Disinfectants and Disinfection Byproducts, reducing the MCL for the four
THMs (THM4: TCM, BDCM, DBCM and TBM) to 80 µg/L and including contaminant
level for the sum of the five HAAs (HAA5: MCAA, MBAA, DCAA, DBAA and
TCAA) at 60 µg/L. Despite all these efforts, these regulations are believed to not be
strong enough and one study of Singer (2006) entitled “Regulation of Only Five
Haloacetic Acids is Neither Sound Science Nor Good Policy” expressed the confusion
of regulating only five HAAs and illustrated situations where overall HAA
concentration and HAA exposure have been under-estimated and have led to confusion
and incorrect interpretation of data. UK regulations have focused so far on THMs and
state a maximum level of 100 µg/L with a frequency of sampling dependant of the
population size (DWI, 2005). In regards to the HAA, the European Union is considering
regulating the nine HAAs at 80 µg/L (Cortvriend, 2008).
17
2.3.2
Analytical measurement
HAAs
HAAs are a group of anions that can be found at levels significantly above the US EPA
regulatory limits and therefore to monitor their level in drinking water, several
analytical methods have been developed:

Gas chromatography – electron capture detector (GC/ECD),

Gas chromatography – mass spectrometry (GC/MS),

Ion chromatography (IC),

Capillary electrophoresis (CE) and capillary zone electrophoresis (CZE),

High performance liquid chromatography (HPLC) and

Electrospray ionization – mass spectrometry (ESI/MS).
Examples of published limits of detection (LOD) from each method have been collated
(Table 2.3). Their sample preparation and the number of HAA detected have also been
reported.
The GC/ECD method is the most widely applied and to date the US EPA has set four
approved methods: the US EPA Method 552.1 (1992), the US EPA Method 552.2
(1995), the US EPA Method 552.3 (2003) and the Standard Method 6251 (APHA,
1998). In these methods, HAAs are extracted from water samples using either methyl
tert-butyl ether (MtBE) or anion exchange resins, and then converted to their methyl
esters using acidic methanol or diazomethane. The GC/ECD US EPA methods are
typically reliable and accurate with detection limits for the nine HAAs in the low µg/L
range, but on the other hand, they involve labour intensive extraction procedures and the
use of toxic derivatisation reagents.
William et al. (1997), Scott et al. (2000) and Xie (2001) determined the HAA
concentrations by GC/MS. This method also requires the conversion of HAAs to their
methyl ester forms. The advantages of the method are (1) the fewer interfering peaks
than with GC/ECD, (2) clean baselines and (3) the use of short run time without
compromising the analytical results (Xie, 2001). On the other hand, the method shows a
low sensitivity for brominated species (Xie, 2001).
18
Liu and Mou (2003) were the first to report the IC method, which is a short method due
to the limited sample preparation. The nine HAAs can be detected but the sensitivity of
the method remains poorer than that of the GC/ECD ones.
Other methods reported for the quantification of HAAs in drinking water include CE or
CZE, HPLC and ESI/MS (Skelly, 1982; Nair et al., 1994; Vichot and Furton, 1994; Xie
and Romano, 1997; Carrero and Rusling, 1999; Urbansky, 2000; Xie et al., 2000; Kim
et al., 2001). CE, CZE and HPLC are short methods (less than 10 minutes to separate
HAA9), but they allow detection only in the mg/L range or in medium µg/L range. On
the other hand, ESI/MS is a very sensitive and selective method that achieves detection
limits of ≤ 70 ng/L. However, the cost of the instrumentation contributes to its lack of
availability in research laboratories.
Concluding comments
HAA levels expected to be in the mg/L range, such as in biodegradation work, will be
easily analysed with HPLC or IC (Hozalski et al., 2001; McRae et al., 2004). However,
the concentration of HAAs in drinking and/or treated water is expected to be in the low
µg/L range (trace levels) and hence, the use of GC/ECD or GC/MS is more suitable
than IC, CE or HPLC.
19
Table 2.3 Analytical method, description, and limit of detection
Measuring
device
GC/ECD
Sample Preparation
Extensive preparation
Derivatisation (extraction +
methylation)
Instrumental
analysis time
GC/MS
IC
CE
20
Moderate preparation
Filatration through a 0.45 µm
pore size filter capsule
HAA detected
0.820 – 0.066 (84.7 – 107)
HAA9
0.012 – 0.170 (97.8 - 108)
HAA9
0.4 – 1.3
HAA6 (BDCAA
instead of MCAA)
0.6 – 1.3
HAA6
0.066 – 0.820
MBAA, DCAA,
TCAA, DBAA, TBAA
NRa (52 – 105)
HAA9
0.07 – 0.83 (73 – 165)
HAA9
0.001 – 0.050
MCAA, MBAA,
DCAA, DBAA
Scott et al. (2000) – Canada
0.37 – 31.6 (96.6 – 99.1)
HAA9
Liu and Mou (2003) –
China
683 – 1507
HAA9
Hozalkski et al. (2001) – US
140 – 1062
MCAA, MBAA,
TCAA
McRae et al. (2004) – US
> 55 min
US EPA method protocols
Derivatisation (extraction +
methylation)
US EPA method protocols +
Method from Scott and Alaee
(1998)
Minimal preparation
Cartridges to remove chlorine
and silver
Limit of detection (µg/L)
(recovery, when applicable,
%)
Reference
US EPA method 552.2
(1995) - US
US EPA method 552.3
(2003) - US
Malliarou et al. (2005) – UK
Rodriguez et al. (2004) –
Canada
Marhaba and Van (2000) –
US
Pourmoghaddas et al. (1993)
– US
Xie (2000) – US
< 55 min
< 40 min
< 10 min
Sample Preparation
Instrumental
analysis time
Limit of detection (µg/L)
(recovery, where applicable,
%)
HAA detected
Reference
HPLC
Skelly’s procedure (1982)
< 10 min
30 – 40
DCAA, DBAA,
TCAA, TBAA
Heller-Grossman et al.
(1993) – Israel
HPLC/EC
Direct evaporation and solid
phase extraction (SPE)
< 30 min
120 – 10000
HAA6 (TBAA instead
of DBAA)
Carrero and Rusling (1999) –
US
ESI/MS
Extraction as for GC/ECD
< 10 min
≤ 70 ng/L
NRa
Urbansky (2000) – US
Measuring
device
a
Not reported.
21
THMs
THMs are volatile and they can be “stripped” from the water into GC gas phase. In the
US there are three approved methods for THM analysis (US EPA, 1995a, 1995b,
1995c). Both US EPA 502.2 (1995b) and 524.2 (1995c) are purge and trap (P&T) GC
methods. US EPA method 551.1 (1995a) uses liquid-liquid extraction (LLE) before
analysis by GC/ECD. There are also three Standards Methods for the analysis of THMs
(6232B, 6232C and 6232D) (APHA, 1998), which are similar to the US EPA ones but
they are not listed as approved methods (Xie, 2003).
2.3.3
Typical levels in drinking water
HAAs and THMs are by far the most abundant DBPs in drinking water and HAAs can
be found equal to, or greater than the concentration of THMs (Singer, 2002). The range
of HAA and THM worldwide concentrations reported in drinking water have been
summarised (Table 2.4).
HAAs and THMs have been extensively studied in the US, where several national
occurrence studies have been undertaken (Krasner et al., 1989; McGuire et al., 2002;
Weinberg et al., 2002; Krasner et al., 2006). For example, Krasner et al. (1989), in their
survey encompassing 35 water treatment utilities, found median total THM
concentration ranged from 30 to 44 µg/L, with TCM, BDCM, DBCM and TBM from
9.6-15, 4.1-10, 2.6-4.5 and 0.33-0.88 µg/L respectively and median total HAA5
concentration ranged from 13 to 21 µg/L, with MCAA, MBAA, DCAA, TCAA and
DBAA from <1-1.2, <0.5, 5.0-7.3, 4.0-6.0 and 0.9-1.5 µg/L respectively. Recently
another survey by Krasner et al. (2006) reported median average HAA9 concentration,
in 12 differently treated drinking waters, to be 34 µg/L and 31 µg/L for median average
THM concentration from the same utilities.
In the UK, one study reported levels of HAAs and THMs on a regional basis, with HAA
means ranging from 35 to 95 µg/L per region (Table 2.4) (Malliarou et al., 2005).
In most of the surveys cited (Table 2.4), DCAA, TCAA and TCM are the most
important species, with DCAA and TCAA similar to the concentrations of TCM.
Furthermore, HAA and THM concentrations have been shown to be affected by
22
seasonal variation with the highest concentrations found in summer (Krasner et al.,
1989; Singer et al., 1995; Williams et al., 1997; Dojlido et al., 1999).
Table 2.4 Reported levels of HAAs and THMs in drinking and/or treated water
THM4
Range
(µg/L)
NR-76
Location
Water source
Malliarou et al.
(2005)
Nissinen et al.
(2002)
Peters et al.
(1991)
Cancho et al.
(1999)
Dojlido et al.
(1999)
UK
Drinking water
Finland
Drinking water
HAA6
6.00-261
1.00-103
Netherlands
Drinking water
HAA9
0.00-14.7
NRa
Spain
Drinking water
HAA9
11.0-32.0
58.0-91.0
Poland
After
chlorination
HAA6
10.0-120
2.00-110
Williams et al.
(1997)
Canada
Drinking water
MCAA
MBAA
DCAA
TCAA
DBAA
0.30-10.0
0.01-9.00
0.20-163
0.04-473
0.01-2.00
0.30-342
Krasner et al.
(1989)
Krasner et al.
(2006)
Ates et al. (2007)
US
Drinking water
HAA5
13.0-21.0
30.0-44.0
US
Drinking water
HAA9
5.00-130
4.00-164
Turkey
Filtered surface
water
Drinking water
HAA9
6.00-177
13.0-191
HAA6
0.40-14.0
3.00-16.0
Wang et al.
(2007)
a
HAA
HAA
Range
measured
(µg/L)
HAA9
NRa-244
Reference
China
Not reported.
Several studies looked at the correlation between HAAs and THMs (Nissinen et al.
2002; Liang and Singer, 2003; Malliarou et al., 2005; Ates et al., 2007). For example
Ates et al. (2007) found a strong correlation between HAAs and THMs with a
coefficient of correlation (R2) of 0.87, whereas Nissinen et al. (2002) found an R2 of
0.35 (Figure 2.3). In their study, Malliarou et al. (2005) found an R2 of 0.56 for one of
the regions they investigated. However, they also reported that in another region, no
correlation could be found between HAAs and THMs. A good correlation between
THMs and HAAs can be useful for quality control and monitoring in water utilities
because, in general, laboratory analyses for HAAs cost considerably more and are also
more time consuming than the THM analyses (Sérodes et al., 2003).
23
200
180
Ates et al. (2007)
y = 1.28x + 3.01
R2 = 0.87
THM concentration (µg/L)
160
140
120
100
Malliarou et al. (2005)
y = 0.21x + 7.67
R2 = 0.56
80
60
40
Nissinen et al. (2002)
y = 0.21x + 4.85
R2 = 0.35
20
0
0
40
80
120
160
200
HAA concentration (µg/L)
240
280
Figure 2.3 Correlation between HAAs and THMs
2.3.4
Influence of organic precursors
2.3.4.1 Water characteristics: TOC/DOC, UV254 and SUVA254
NOM consists of humic (non-polar, hydrophobic) and non-humic (polar, hydrophilic)
substances, generally of terrestrial and biological origin respectively (Hwang et al.,
2001). NOM provides the precursor material from which DBPs are formed; hence, the
amount (described as DOC or TOC) and the nature (described as UV254) of NOM will
give some insights into the DBPs formed. SUVA (expressed in L/m-mg C) is defined as
the UV absorbance (1/m) of a given sample determined at 254 nm and divided by the
DOC concentration (mg/L) of the sample, and can give information on the type of NOM
present in the sample. Guidelines for SUVA are shown below (Table 2.5).
Table 2.5 Guidelines for the nature of NOM (Edwald and Tobiason, 1999)
SUVA
(L/m-mg C)
>4
2-4
<2
24
Composition of the water source
Mostly aquatic humics.
High hydrophobicity, high molecular weight (MW)
Mixture of aquatic humics and other NOM.
Mixture of hydrophobic and hydrophilic NOM, intermediate MW
Mostly non-humics.
Low hydrophobicity, low MW
TOC/DOC, UV254 and SUVA254 are surrogate parameters for estimating the extent of
DBP formation. A review of the literature has identified a number of relationships
between the water characteristics, HAAs and THMs.
First of all, data from a range of 43 drinking and treated waters from 5 references
(Allgeier and Summers, 1995; Ratnaweera et al., 1999; Nokes et al., 1999; Siddiqui et
al., 2000; Volk et al., 2000) and 32 raw waters from 6 references (Collins et al., 1986;
Allgeier and Summers, 1995; Ratnaweera et al., 1999; Vilgé-Ritter et al., 1999; Siddiqui
et al., 2000; Volk et al., 2000) were collated. A good linear relationship between the
concentration of organic matter (expressed as DOC or TOC) and UV254 was observed in
drinking and treated waters with an R2 of 0.96 (Figure 2.4). For comparison, the same
relationship was investigated for raw waters and it was observed a similar R2 (0.94),
showing that the relationship is similar before and after treatment.
Absorbance at 254 nm (1/m)
120
Drinking & treated waters
Raw waters
100
Raw waters
y = 4.56x - 5.82
R2 = 0.94
80
60
Drinking & treated waters
y = 4.32x - 2.97
R2 = 0.96
40
20
0
0
5
10
15
20
Concentration of organic matter (mg/L)
25
Figure 2.4 Relationship between the organic matter concentration (TOC/DOC) and
UV254 for a range of 43 drinking and treated waters and 32 raw waters
Secondly, it has been reported a linear positive relationship between HAA and THM
formation potential (FP) and the water characteristics (Shorney et al., 1999; Hwang et
al., 2001; White et al., 2003). The correlation between DOC and HAA-FP and THM-FP
25
were good, with an R2 above 0.73 (Table 2.6). Ates et al. (2007) found an exponential
correlation between HAAs, THMs and DOC, with an R2 of 0.88 and 0.92 respectively.
Although excellent correlations have been reported between DOC and DBP-FP for a
single water, these correlations are not as good when waters from different sources are
included, and this because waters from different sources tend to have different specific
DBP yields as determined by their particular watershed characteristics (Reckhow and
Singer, 1990). A Combination of good and moderate relationships was reported
between DBPs and UV254 and SUVA254 (Table 2.6).
Table 2.6 Relationship between HAA-FP and THM-FP and water characteristics in
treated waters
Number of water samples
2
R for HAA-FP to DOC
R2 for THM-FP to DOC
R2 for HAA-FP to UV254
R2 for THM-FP to UV254
R2 for HAA-FP to SUVA254
R2 for THM-FP to SUVA254
a
White et al.
(2003)
Shorney et
al. (1999)
Ates et al.
(2007)
Hwang et
al. (2001)
15
19
29
7
0.86
0.87
0.99
0.99
0.90
0.91
0.73
0.83
0.59
0.92
0.71
0.55
a
0.88
0.92 a
0.91
0.93
0.77
0.69
0.80
0.97
0.91
1.00
NR b
NR b
Exponential relationship; b Not reported.
Reckhow et al. (1990) also reported a linear positive relationship between THM-FP and
SUVA254. A linear positive relationship has also been reported for HAA-FP and
SUVA254 with higher correlations than those shown for THM-FP (Singer et al., 2002).
Neither the relationship reported between both DBPs nor SUVA254 and HAA-FP was as
strong as the relationship reported between both DBPs and TOC/DOC and UV254
separately (Singer et al., 2002).
2.3.4.2 NOM fractionation
NOM found in water consists of both hydrophobic and hydrophilic species that
contribute considerably to the formation of DBPs (Goslan et al., 2002). Typically, the
hydrophobic is the largest fraction of the NOM pool (Hwang et al., 2001). In order to
better understand the link between precursors and DBPs, the different fractions of
aquatic NOM are isolated and fractionated and then chlorinated to determine their DBPFP. Eight fractions have been reported in the literature:
26

Hydrophilic base (HPI-B): amphoteric proteinaceous materials containing amino
acids, amino sugars, peptides and proteins (Leenheer, 1981),

Hydrophilic acid (HPI-A): organic compound of the hydroxyl group (Leenheer,
1981),

Hydrophilic neutral (HPI-N): organic compound made up of polysaccharides
(Marhaba and Van, 2000),

Colloids: consistent composition of predominantly non-reactive carbohydrates
and N-acetylaminosugars (Hwang et al., 2001),

Hydrophobic base (HPO-B): humic substance (Leenheer, 1981),

Hydrophobic acid (HPO-A): a soil fulvic (Marhaba and Van, 2000),

Hydrophobic neutral (HPO-N): a mix of hydrocarbon and carbonyl compounds
(Leenheer, 1981),

Transphilic (TPI): intermediate polarity, generally more hydrophobic character
than hydrophilic, but it is variable and highly dependent on the particular source
of the transphilic NOM (Hwang et al., 2001).
In the UK, chlorination tends to occur after the water has been treated by a coagulant
and filtered (Parsons and Jefferson, 2006; Sharp et al., 2006), which is different from
US treatment practices where pre-chlorination is widely used (Singer et al., 2002). After
conventional treatment, NOM is mainly hydrophilic in character and low in
concentration (Goslan et al., 2002). However, a review of the literature undertaken on
drinking and treated waters or waters with low humic content (Table 2.7) showed that
the hydrophilic NOM can contribute substantially to the formation of DBPs.
Kanokkantapong et al. (2006) found the hydrophilic neutral fraction to be the most
reactive towards the formation of HAAs, whereas Marhaba and Van (2000) reported the
hydrophobic neutral to be the most reactive fractions for the formation of HAAs and
THMs. Sinha et al. (1997) and Goslan et al. (2002) reported the humic (hydrophobic)
fraction to be very reactive in regards to the THMs. Kim and Yu (2005) found the
hydrophobic and the hydrophilic fractions to give similar THM concentrations, and the
27
hydrophilic fraction to be the most reactive towards the formation of HAAs. Croué et al.
(2000) reported similar DBP concentrations for each fraction studied.
Two studies reported levels of chlorine demand from each fraction (Table 2.7). No
specific trend could be observed. Kanokkantapong et al. (2006) reported the hydrophilic
neutral to be the most reactive fraction with chlorine, whereas Croué et al. (2000) found
the transphilic neutral as the most active fraction.
28
Table 2.7 Fractions, their characteristics, their chlorine demand and their DBP-FP
Fractions
Hydrophobic neutral
Hydrophobic base
Hydrophobic acid
Hydrophilic base
Hydrophilic acid
Hydrophilic neutral
Humic acid
Fulvic acid
Hydrophilic acid
Hydrophilic non acid
Hydrophobic
Hydrophilic
Hydrophobic neutral
Hydrophobic base
Hydrophobic acid
Hydrophilic base
Hydrophilic acid
Hydrophilic neutral
Humic acid
Non-humic acid
Treatment
Distribution of
DOC (%)
Filtered water
5.0 – 12
1.0 – 3.0
30- 33
3.0 – 5.0
8.0 – 19
31 – 41
Chlorine
demand
(mg/L)
0.61
0.22
1.33
1.33
0.40
3.47
2.0 – 5.0
14 – 33
1.0 – 11
54 – 80
NR a
NR a
NR a
NR a
35
65
NR a
NR a
Effluent water
8.6
11
11
11
40
17
NR a
NR a
NR a
NR a
NR a
NR a
Optimized
coagulation
NR a
NR a
NR a
NR a
Coagulation,
dissolved air
flotation, rapid
gravity
filtration
Filtered water
HAA
detected
HAA-FP
(µg/mg C)
THM-FP
(µg/mg C)
HAA5
3.52
5.72
15.2
4.72
6.78
28.5
NR a
NR a
NR a
NR a
NR a
NR a
Correlation
with DOC
(R2)
0.98
1.00
1.00
1.00
0.96
1.00
NR a
NR a
NR a
NR a
11.7 – 154
84.3 - 92.0
11.9 - 43.9
4.00 - 70.2
NR a
NR a
NR a
NR a
Goslan et al.
(2002)
16.3
24.1
42.3
41.8
NR a
NR a
Kim and Yu
(2005)
HAA6
2.00
0.60
0.80
0.50
NR a
NR a
2.00
0.40
1.20
0.30
10.0
NR a
NR a
NR a
NR a
NR a
NR a
NR a
HAA6
40.0 - 100
40.0 – 60.0
45.0 - 150
25.0 – 75.0
NR a
NR a
NR a
NR
a
Reference
Kanokkantapong
et al. (2006)
Marhaba and Van
(2000)
Sinha et al.
(1997)
29
Fractions
Hydrophobic acid
Hydrophobic neutral
Hydrophobic base
Hydrophilic
Hydrophobic neutral
Hydrophobic acid
Transphilic acid
Transphilic neutral
Hydrophilic acid
Hydrophilic neutral
Hydrophobic
Transphilic
Hydrophilic
acid+neutral
Hydrophilic base
Colloids
a
Treatment
Distribution of
DOC (%)
Filtered water
47
30
16
7.0
Filtered water b
Ozone
contractor
effluent c
3.0
31
14
12
Chlorine
demand (mg
Cl2/mg C)
NR a
NR a
NR a
NR a
6.0
0.83
0.95
0.81
2.30
0.86
1.00
36
15
NR a
NR a
7.5
NR a
0.2
0.5
NR a
NR a
HAA
detected
HAA-FP
(µg/mg)
THM-FP
(µg/mg)
HAA5
142
40.0
20.0
120
135
100
120
60.0
Correlation
with DOC
(R2)
NR a
NR a
NR a
NR a
TCAA
and
DCAA
28.0
42.0
35.0
32.0
40.0
34.0
29.0
46.0
39.0
25.0
35.0
28.0
NR a
NR a
NR a
NR a
NR a
NR a
0.04 d
0.06 d
0.07 d
0.09 d
NR a
NR a
0.21 d
0.17 d
NR a
0.17 d
0.06 d
0.12 d
0.04 d
NR a
NR a
HAA9
Not reported; b DOC loss = 34%; c DOC loss = 41%; d DBP yield (µmol/µmol) expressed in percentage (%).
30
Reference
Chang et al.
(2001)
Croué et al.
(2000)
Hwang et al.
(2000)
In their study investigating polar NOM, Hwang et al. (2001) reported the hydrophilic
base to be the most reactive fractions (Table 2.8) towards chlorine and the hydrophilic
acid + neutral to be the most reactive fraction in regards to the formation of HAAs and
THMs. They also reported that the ratio between HAAs and THMs was different in
each fraction and the DXAAs were always greater in each fraction than the TXAAs
(Table 2.8). A summary of the fraction composition is also given (Table 2.8).
Table 2.8 Summary of NOM properties (Hwang et al., 2001)
Fraction/
Hydrophobic
Parameter
Cl2 demand
+
THM yield
+++
HAA yield
++
HAA9/THMs THMs>HAA9
DXAA/TXAA
~1.1
ratio
Molecular
Highest
weight
DBPprecursor
Ketones
(13C NMR)
Aromatic C=O
Phenols
Composition
(pyrolysis)
Aromatic
Phenolic
Transphilic
+
+++
++
THMs>HAA9
~1.6
Intermediate
Hydrophilic
Hydrophilic
Colloids
A+N d
Base
++
++++
+
++++
++
+
++++
variable
+
THMs~HAA9 HAA9>THMs THMs~HAA9
~2.4
~2.7
~2.9
Lowest
NA a
NA a
Less aromatics,
NA a
Hydroxyl acids, Hydroxy acids
Methylenes
Carbodydrates
Aromatic
Aminosugars
Proteins and
Phenolic
di-, tri-, mixed
aminosugars
Proteins
alcohols and
acids
+ Lowest yield, ++++ highest yield (µmol DBP/µmol DOC); a Not analysed;
Polyhydroxybutyrates; d Hydrophilic acid + neutral.
b
NA a
Carbodydrates
Aminosugars
DNA b,
PHBs c, fatty
acids
Deoxyribonucleic acid;
c
Whilst a review of the literature showed significant contradictions in which fractions
contribute the most to the DBP formation, it is true that the hydrophilic material can
contribute to the formation of DBPs. Therefore, NOM, that is mainly hydrophilic in
character after conventional treatment processes, should be studied for their DBP-FP
(Goslan et al., 2002).
2.3.5
Influence of inorganic precursors: bromide
When chlorine is added to the water containing bromide, the bromide ions are oxidised
to hypobromous acid as shown in Equation 2.4, and hypobromous acid reacts with
NOM to form brominated DBPs. Heller-Grossman et al. (1993) and Cowman and
31
Singer (1996) reported that low bromide-containing waters disinfected with chlorine
formed preferentially TCM, DCAA and TCAA, whereas there were a dominance of
brominated DBPs in chlorinated waters containing levels of bromide > 100 µg/L.
Typical concentrations of bromide in natural waters usually range from 30 to 200 µg/L,
with an average of 100 µg/L (Amy et al., 1994). Hence, it is likely that most of the
waters disinfected with chlorine will have the potential to form brominated DBPs.
Hua et al. (2006) observed the shift of DBP speciation towards the brominated species
when adding bromide into the water sources. Indeed they reported that the total
concentration of the regulated HAA5 (MCAA, MBAA, DCAA, TCAA and DBAA)
decreases as the bromide concentration increases because DCAA and TCAA
significantly decrease and only two brominated species are measured, whereas the
addition of bromide increases the total HAA9 (Table 2.9). An investigation into the
effect of bromide ions on HAA formation reported that bromine is more reactive than
chlorine in the reactions of substitution and addition that form HAAs (Cowman and
Singer, 1996). Specifically, in halogen substitution of DXAA formation, hypobromous
acid is 25 times stronger than free chlorine (hypochlorous acid) (Chang et al., 2001).
Table 2.9 Effect of bromide on the molar yield of THMs and HAAs (reaction
conditions: pH = 7, contact time = 48 h, temperature = 20°C, Cl2 dose = 6.2 mg/L and
initial water source bromide concentration = 63 µg/L) (Hua et al., 2006)
Bromide addition
(µmol/L)
0
2
10
30
HAA5
0.90
0.70
0.40
0.50
DBPs (µmol/L)
HAA9
1.05
1.10
1.25
1.40
THM4
1.25
1.45
1.95
2.20
In terms of THMs, Hua et al. (2006) reported that increasing initial bromide
concentrations resulted in a substantially increase of 74% of the THM molar
concentration (Table 2.9). Amy et al. (1991) found that free chlorine is a more effective
oxidant, whereas hypobromous acid behaves as a more efficient halogen substitution
agent in the formation of THMs. Hence, more than 50% of the bromide is incorporated
into THMs compared to 5 to 10% of the chlorine. Additionally hypobromous acid
attacks more sites in the THM precursors and reacts with them faster than with free
chlorine (Krasner, 1999).
32
It should be noted that while doing FP tests, chlorine is used in excess leading to more
chlorinated DBPs being formed than brominated ones (Symons et al., 1993).
2.3.6
Influence of water treatment variables
2.3.6.1 pH
It is well established that the formation of THMs increases with increasing pH (Singer,
1999; Kim et al., 2002; Xie, 2003) as it has been shown that the formation of THMs
consists of alternate hydrolysis and halogenation steps that are enhanced at higher pH
(Trussell and Umphres, 1978). However, the effect of pH on the formation of HAAs is
equivocal.
Overall, HAA formation increases with decreasing pH (Krasner, 1999). Specifically,
Liang and Singer (2003) reported that increasing pH from 6 to 8 had a little effect on the
formation of MXAA and DXAA, but significantly decreased the formation of TXAA
and in particular TCAA. DCAA, the other species the most affected by pH, was
reported to be the highest at pH 7 on a pH ranges from 5.0 to 9.4 (Krasner et al., 1999).
The speciation of HAAs and THMs at different pH values is determined by the
formation mechanisms of the different species. Based on Reckhow and Singer’s
mechanism (1985), THMs and TXAA have a common precursor structure (R-CO-CX3),
and the relative formation of these species is determined by the nature of the R group
and pH. Under alkaline conditions, base-catalysed hydrolysis prevailed, yielding more
THMs. In acidic environments, on the other hand, TXAA will be formed if the R group
is a readily oxidisable functional group capable of easily donating an electron pair to the
rest of the molecule. In the absence of such an oxidative cleavage (e.g., if the R group is
not a readily oxidisable functional group), hydrolysis might still prevail, resulting in
THMs (Singer et al., 2002). The model of Reckhow and Singer (1985) also showed that
there might be more precursor structures and formation pathways for DXAA than for
TXAA, which may make the formation of DXAA exhibit more complex behaviour with
respect to pH (Singer et al., 2002).
33
2.3.6.2 Temperature
The temperature has been previously correlated with HAAs by Dojlido et al. (1999),
who reported that during the winter season (1°C) the level of HAAs was 0.63 µg/mg C,
whereas during the summer (23°C), concentration reached 7.4 µg/mg C. In a study by
Malliarou et al. (2005), temperature was significantly correlated with the ratio of total
THM and total HAA.
However, the effect of temperature (rise from 0 to 33°C) on the THM yield was rather
limited (El-Dib and Ali, 1995). Specifically, Carlson and Hardy (1998) found that
THMs were impacted by the temperature, but this at longer contact time.
It is thought that the nature of NOM may also differ from summer to winter and could
be responsible for the difference of DBP concentration measured.
2.3.6.3 Alternative disinfectants
As HAAs and THMs are formed by the reaction of NOM with chlorine, methods of
control include the use of alternative disinfectants, such as chloramines.
Chloramines
include
monochloramine
(NH2Cl),
dichloramine
(NHCl2)
and
trichloramine (NCl3) (Section 2.2.2). However, monochloramine is the predominant
chloramine species present under the controlled conditions typically found in water
treatment works and distribution systems (Vikesland et al., 1998). Many utilities have
switched to monochloramine as a secondary disinfectant because it generally results in
lower concentrations of THMs and HAAs (Speitel, 1999; Diehl et al., 2000). With
concurrent addition of chlorine and ammonia, the HAA formation is typically 5 to 20%
of that observed with chlorine alone (Speitel, 1999) and DXAAs are the most
commonly formed DBPs comprising > 90% of the total HAA (Diehl et al., 2000). Many
plants have a significant period of exposure to free chlorine before ammonia addition
for purpose of meeting CT requirements (Section 2.2.2) (Seidel et al., 2000). This
prechlorination period, necessitated by disinfection regulation, is of greatest concern
when DBP formation is considered as it gives free chlorine the opportunity to react with
NOM. Clearly, by adding free chorine first rather than adding monochloramine, more
HAAs and THMs are formed. Pope et al. (2006) looked at the effect of prechlorination
34
and concluded that during short periods of prechlorination (5 or 20 minutes),
significantly more DXAA formation, as well as MXAA and TXAA, occurred, relative
to the period of free chlorination. In one of their waters studied (Pope et al., 2006), the
use of preformed chloramines resulted in 9 µg/L of DXAA after 48 hours incubation,
whereas prechlorination for 5 and 20 minutes formed 25 and 35 µg/L of DXAA
respectively. The effect of pH and Cl2:N ratio impacts also the formation of DBPs, and
by increasing the Cl2:N ratio and decreasing the pH, HAA formation will be maximised
(Diehl et al., 2000). For example, Diehl et al. (2000) reported that a pH of 6 and a Cl2:N
ratio of 7:1 resulted in about 19 µg/L of HAAs formed, whereas in the same water, at a
pH 10 and a Cl2:N ratio of 3:1, less than 5 µg/L of HAAs were formed.
When monochloramine reacts with NOM to form DBPs, the DBP formation pathways
remain unclear. When bromide is absent, monochloramine and its two decomposition
products, free chlorine and dichloramine potentially react with NOM to form HAAs
(Karanfil et al., 2008). Hong et al. (2007) and Karanfil et al. (2007) showed that the
direct reaction between monochloramine and NOM plays the major role and is
responsible for about 80% of HAA formation and that the remaining HAA formation
was attributed to free chlorine that results from monochloramine decomposition. No
HAA formation was attributed to dichloramine. Whilst Duirk et al. (2005) and Duirk
and Valentine (2006) concluded that DXAA formation was most due to the direct
reaction of free chlorine, in quasi-equilibrium with monochloramine, with NOM; a
schematic showing the pathways are presented below (Figure 2.5).
NH2Cl
DOC1
NH4+ + Cl-
DOC2
DBPs
NH3 + HOCl
Figure 2.5 Schematic of monochloramine reaction pathways in the presence of NOM
(as DOC). Monochloramine reacts directly with reactive site concentration [DOC1], and
free chlorine with reactive site concentration [DOC2]. Adapted from Duirk and
Valentine (2006)
35
The presence of bromide significantly complicates the chemistry (Diehl et al., 2000;
Karanfil et al., 2007). Bromamines, chloramines and bromochloramine would co-exist
in a chloraminated samples and Pope et al. (2006) found that the dihalogenated
bromine-substituted species (BCAA and DBAA) formed more rapidly than the DCAA
because bromochloramine has a higher reactivity than monochloramine (Vikesland et
al., 2001).
Other alternative disinfectants include ozonation and chlorine dioxide. Ozonation is
used in the UK as a pre-treatment but is usually associated to the formation of bromate
(Jarvis et al., 2007). Although chlorine dioxide is used in the US, this is not the case in
the UK because of the chlorite/chlorate by products associated to its use. DWI (2009)
states the level of chlorine dioxide, chlorate and chlorite permitted in drinking water to
be 0.5 µg/mL in total.
2.3.6.4 Disinfectant dose
Increasing the chlorine dose results in increased DBP formation up to a point where the
concentrations reached an equilibrium (Fleischacker and Randtke, 1983; Carlson and
Hardy, 1998). Fleischacker and Randtke (1983) stated that after 96 hours contact time, a
dose less than 6 mg free chlorine per mg DOC, THMs form in proportion to increasing
free chlorine, whereas at a dose greater than 6 mg free chlorine per mg DOC, little
additional THMs form as free chlorine increases. Carlson and Hardy (1998) also found
that the point at which the reaction was no longer chlorine limited was 1 mg free
chlorine per mg of TOC; this evaluated at a shorter contact time (< 7 hours). However,
it is thought that the nature of NOM in the water will also have an effect on the point at
which the reaction is no longer chlorine limited. Similar patterns were also observed for
HAAs.
36
2.3.6.5 Reaction time and kinetics
The effect of the reaction time has been widely studied with the common statement that
HAAs and THMs were formed rapidly in the first few hours of the reaction and then the
formation slowed as the concentrations of the reactants (either NOM or chlorine)
decreased with time (Reckhow et al., 1990; Singer, 1999; Gallard and Von Gunten,
2002; Singer et al., 2002; Liang and Singer, 2003; Nikolaou et al., 2004).
From this finding, the fast rate of reaction has been referred to as a first order reaction
and the slower rate, second order. The reactions and the calculation of the rate are
shown below:
1st Order : A  B  C
Rate = k1[A]1[B]1=k[A][B],
Equation 2.8
2 nd Order : A  2B  C
Rate = k2[A]1[B]2=k[A][B]2,
Equation 2.9
where k1 and k2 are the constant of 1st and 2nd order respectively, A and B are reactants
(NOM and disinfectant) and C is the products (DBPs) (Chowdhury and Amy, 1999).
2.3.7
Predictive models for HAAs and THMs
Many models have been developed to predict THM and HAA formation (Amy et al.,
1987; Adin et al., 1991; Lou and Chiang, 1994; Chang et al., 1996; Rathbun, 1996;
Garcia-Villanova et al., 1997; Clark, 1998; Chowdhury et al. 1999; Rodriguez et al.,
2000; Golfinopoulos and Arhonditsis, 2002; Gang et al., 2003). The creation of
predictive models requires a large database of existing results. Two studies carried out
by Sohn et al. (2004) and Sadiq and Rodriguez (2004) reviewed the existing models for
the prediction of DBPs in raw, drinking and treated water. The review by Sohn et al.,
(2004) focused on several types of models, such as empirical power function and
empirical kinetic models, in order to find more about the robustness and the accuracy
for the DBP prediction. What has been concluded from these reviews is the greater
amount of models available for THMs in comparison with those available for HAAs.
Only 22% of the models reviewed by Sadiq and Rodriguez (2004) are related to HAAs.
A summary of selective predictive HAA models for application in raw and treated
waters are presented (Table 2.10). The HAA models identified in this summary are of
37
the form: DOC/UV based models (Watson, 1993; Amy et al., 1998; Sohn et al., 2004),
chlorine demand models (Gang et al., 2002; Gang et al., 2003) and linear regression
models (Villanueva et al, 2003; Sérodes et al., 2003).
38
Table 2.10 Selected HAA predictive models
Species
(µg/L)
Predictive models for HAA
MCAA 1.634 (TOC)0.753 (Br- + 0.01)-0.085 (pH)-1.124 (Cl2)0.509 (t)0.300
DCAA 0.605 (TOC)0.291 (UV)0.726 (Br- + 0.01)-0.568 (Cl2)0.48 (t)0.239 (T)0.665
TCAA 87.182 (TOC)0.355 (UV)0.901 (Br- + 0.01)0.679 (pH)1.732 (Cl2)0.881 (t)0.264
MBAA 0.176 (TOC)1.664 (UV)-0.624 (Br-)0.795 (pH)-0.927 (t)0.145 (T)0.45
DBAA 84.945 (TOC)-0.62 (UV)0.651 (Br-)1.073 (Cl2)-0.2 (t)0.12 (T)0.657
HAAs
HAAs
Linear regression in function of various THM species
R2
Water
source
Advantages
Limitations
Reference
The water quality
and the
0.97
chlorination
Use of
conditions do not
0.98 Not real
independent
Watson
represent
water
databases for
(1993)
0.8
situations
model validation
encountered in
0.95
real water
utilities
Data and models
have potential
Models do not
applications for
consider
chlorine Villanueva et
0.57 –
exposure
NR a
dose,
0.97
al. (2003)
assessment in
temperature, etc.
epidemiological
studies
0.82
Single linear and non-linear regression models for water of single 0.56 –
utility
0.92
NR a
Can represent the
Variation
real seasonal
according to the
variations of
DBP species to
environmental
be modelled and
and water quality
to the utility
characteristics
Sérodes et al.
(2003)
39
Species
(µg/L)
HAA6
HAA6
DOC-based model
HAA6 = 9.98 (DOC)0.935 (Cl2)0.443 (Br-)-0.031 (T)0.387 (pH)-0.655 (t)0.178
UV-based models
HAA6 = 171.4 (UV)0.584 (Cl2)0.398 (Br-)-0.091 (T)0.396 (pH)-0.645 (t)0.178
DOC*UV-based models
HAA6 HAA6 = 101.2 (DOC*UV)0.452 (Cl2)0.194 (Br-)-0.0698 (T)0.346 (pH)-0.623
(t)0.180
HAA6
HAA6
HAA6
40
R2
Water
source
Advantages
Limitations
Amy et al.
(1998)
0.87 Raw water
0.80 Raw water
Reference
Strength of the
model:
flexibility and
applicability.
To a various water
quality as well as
different operating
conditions
Sohn et al.
(2004)
Sohn et al.
(2004)
0.85 Raw water
DOC-based model
HAA6 = 5.22 (DOC)0.585 (Cl2)0.565 (Br-)-0.031 (t)0.153
Coagulated
water
0.92
(alum or
iron)
Simplicity,
flexibility and
applicability
Do not take into
account pH and
temperature. Use of
another equation for
correction
Amy et al.
(1998)
UV-based models (developed from EPA 1998 database)
HAA6 = 63.7 (UV)0.419 (Cl2)0.640 (Br-)-0.066 (t)0.161
Coagulated
water
0.92
(alum or
iron)
Simplicity,
flexibility and
applicability
Do not take into
account pH and
temperature. Use of
correction
Sohn et al.
(2004)
DOC*UV-based models (developed from EPA 1998 database)
HAA6 = 30.7 (DOC*UV)0.302 (Cl2)0.541 (BR-)-0.012 (t)0.161
Coagulated
water
0.94
(alum or
iron)
Simplicity,
flexibility and
applicability
Do not take into
account pH and
temperature. Use of
correction
Sohn et al.
(2004)
Species
(µg/L)
HAA6
pH and temperature correction
HAA6 = (HAA6@pH = 7.5, T = 20°C)*(0.932)(pH-7.5) (1.021)(T-20)
HAA5 4.8 * 104 [OH-]0.35 (C0(1-exp(-kt))0.43)(UV254)0.34
HAA9
Model based on chlorine demand
HAA9 = ßC0 {1-fe-kR*t - (1-f)e-kS*t}
R2
Water
source
Advantages
Limitations
This equation
modify the
existing
Temperature and
Coagulated
coagulated
pH correction
water
water DBP
0.85
factors applicable in
(alum or
models, so that
coagulated waters
iron)
there are
only
applicable under
different pH and
temperature
Include
temperature,
Concentration of
[OH-] calculated
pH, chlorine
0.74 Raw water
dosage and
from the raw water
pH and temperature
UV254
absorbance
The model may not
perform well if
tested outside the
The model can
From raw
be applied
typical conditions
(pH 8.0 ± 0.2,
0.98 to treated accurately from
water
raw to alum
temperature = 25°C
treated water
and chlorine
residual = 1.0 ± 0.5
mg/L)
Reference
Sohn et al.
(2004)
Sung et al.
(2000)
Gang et al.
(2002)
TOC/DOC = concentration of NOM (mg/L); Br- = bromide ion (µg/L); Cl2 = chlorine dose (mg/L); t = time in hours; UV = UV absorbance at 254 nm (1/cm); T =
Temperature (°C); ß = HAA9 yield coefficient, defined as the ratio of the concentration (µg/L)of HAA9 formed to the concentration of chlorine consumed (mg/L); C0 =
initial chlorine concentration (mg/L); f = fraction of the chlorine demand attributed to rapid reactions; kR and kS = the first order rate constants for rapid and slow
reactions respectively; a Not reported.
41
Evaluation of models
DOC-, UV- and DOC*UV-based models have been evaluated for their accuracy with
data available in the literature. These models have been chosen because they are the
only models that could fit a large database of existing results. Moreover, because this
study focused on treated water, it was decided to evaluate specific models that fit this
type of water. For comparison, models were also evaluated for raw water. The models
are:
Amy et al. (1998):

DOC-based model in raw water
HAA6 = 9.98 (DOC)0.935 (Cl2)0.443 (Br-)-0.031 (T)0.387 (pH)-0.655 (t)0.178

DOC-based model in coagulated water (alum or iron)
HAA6 = 5.22 (DOC)0.585 (Cl2)0.565 (Br-)-0.031 (t)0.153
Sohn et al. (2004):

UV-based models in raw water
HAA6 = 171.4 (UV)0.584 (Cl2)0.398 (Br-)-0.091 (T)0.396 (pH)-0.645 (t)0.178

UV-based models in coagulated water (alum or iron)
HAA6 = 63.7 (UV)0.419 (Cl2)0.640 (Br-)-0.066 (t)0.161
Sohn et al. (2004):

DOC*UV-based models in raw water
HAA6 = 101.2 (DOC*UV)0.452 (Cl2)0.194 (Br-)-0.0698 (T)0.346 (pH)-0.623 (t)0.180

DOC*UV-based models in coagulated water (alum or iron)
HAA6 = 30.7 (DOC*UV)0.302 (Cl2)0.541 (BR-)-0.012 (t)0.161
In the treated water (coagulated), the temperature and pH are not considered; therefore,
the correction factor (Table 2.10) was applied.
42
The observed HAA concentrations have been fitted against the predicted HAA
concentrations, obtained using the models cited above, in raw and treated waters (Figure
2.6 A+B). Overall results show that the predicted HAA concentrations follow the same
pattern than the observed ones, with more HAA predicted, when there were more HAA
observed. However, in some sources, such as S1 and S7, a difference of more than 1000
µg/L could be observed in raw water, and a difference of more than 200 µg/L in S4,
treated water. The greatest inaccuracy has been observed in water with the highest level
of HAA observed. Therefore, it is believed that these major differences are due to
important parameters such as DOC or UV that remain low for a high HAA formation.
The general trend of these models was to under-estimate the formation of HAAs in both
raw and treated waters. Overall, it seems that, as yet, it is not possible to accurately
predict HAA concentrations with a universal model. The models tested would work
better for certain water sources than for others.
2500
A
2000
Predicted HAAs with DOC-based models in
raw waters (Amy et al., 1998)
1500
1000
Predicted HAAs with UV-based models in raw
waters (Sohn et al., 2004)
500
0
S1 S2 S3 S4 S5 S6 S7 S8
Sources
Observed HAAs in raw waters (Sinha et al.,
1997)
300
B
250
Predicted HAAs with DOC*UV-based model
in raw waters (Sohn et al., 2004)
Observed HAAs in treated waters (Sinha et al.,
1997)
Predicted HAAs with DOC-based model in
treated waters (Amy et al., 1998)
200
150
Predicted HAAs with UV-based model in
treated waters (Sohn et al., 2004)
100
50
Predicted HAAs with DOC*UV-based model
in treated waters (Sohn et al., 2004)
0
S1
S2
S4
S5 S6
Sources
S7
S8
Figure 2.6 Observed HAAs against predicted HAAs using the models of Amy et al.
(1998) and Sohn et al. (2004) in raw (A) and treated waters (B)
43
As a conclusion, scientists have to bear in mind that developing effective and accurate
predictive models for both THMs and HAAs, as well as the rest of the DBPs will allow
a decrease of DBPs measurement studies that require extensive experimental procedures
and chromatographic analysis, which is time and cost consuming. However, developing
accurate model is a complicated and difficult task that requires a considerably large
database (Clark et al., 2001).
2.4
SEMI-VOLATILE DBP OCCURRENCE AND FORMATION FROM
DRINKING WATER
While THM4 are currently measured and regulated in UK waters and HAAs are
considered for regulation, the literature reports more than 600 to 700 DBPs for the most
common disinfectants used (chlorine, chloramines, chlorine dioxide and ozone)
(Richardson et al., 1998; Krasner et al., 2006). Amongst those, 50 DBPs have been
reported as high priority for potential toxicity (Krasner et al., 2001). These DBPs
include i-THMs, HANs, HAs, HKs, HNMs, haloacids, haloamides, MX [3-chloro-4(dichloromethyl)-5-hydroxy-2-(5H)-furanone]
and
its
analogues,
and
N-
nitrosodimethylamine (NDMA) (Mitch and Sedlak, 2002; Richardson et al., 2002;
Weinberg et al., 2002; Richardson, 2003; Krasner et al., 2006).
This literature review will focus on i-THMs, HANs, HAs, HKs and HNMs, because
they are the most common detected in drinking water; they will be referred as semivolatile DBPs. The DBPs that will be further extensively studied in this study are
presented below with their molecular formula, their molecular weight, their structure,
their boiling point and their CAS number (Table 2.11).
44
Table 2.11 Chemical and physical properties of selected semi-volatile DBPs
Semi-volatile DBP
Molecular
formula
Molecular
weight
(g/mol)
Boiling
point
(°C)
CAS
number
110-112
3018-12-0
83-84
545-06-2
NRa
83463-62-1
NRa
3252-43-5
CH3
117 - 118
513-88-2
CH3
NRa
918-00-3
112
76-06-2
NRa
598-91-4
H
88
79-02-7
H
98
75-87-6
NRa
594-04-7
NRa
34970-00-8
Structure
Haloacetonitriles
Cl
DCAN
C2HCl2N
N
C
109.94
Cl
Cl
TCAN
C2Cl3N
144.39
N
C
Cl
Cl
BCAN
C2HBrClN
154.39
Br
N
C
Cl
Br
DBAN
C2HBr2N
N
C
198.84
Br
Haloketones
Cl
1,1-DCP
C3H4Cl2O
126.96
Cl
O
Cl
Cl
1,1,1-TCP
C3H3Cl3O
161.41
Cl
O
Halonitromethanes
Cl
TCNM (also called
chloropicrin)
O-
Cl
CCl3NO2
164.38
N+
Cl
O
Br
DBNM
CHBr2NO2
218.83
ON+
Br
O
Haloaldehydes
Cl
DCA
C2H2Cl2O
112.94
Cl
O
Cl
TCA (also called chloral
hydrate)
Cl
C2HCl3O
147.39
Cl
O
Iodo-Trihalomethanes
Cl
DCIM
CHCl2I
I
210.83
Cl
Cl
BCIM
CHBrClI
I
255.28
Br
a
Not reported.
45
2.4.1
Toxicity and regulations
Many of the semi-volatile DBPs, discovered more recently than THMs and HAAs, were
rated high priority, because they showed a potential toxicity (Richardson, 1998; Krasner
et al., 2001; Plewa et al., 2004). Within the group of halogenated alkanes, such as
THMs and i-THMs, the brominated and the iodinated species are of greater concern
than their chlorinated counterparts, because iodine and bromine are better leaving
groups than chlorine due to their greater polarisable bondings (Woo et al., 2002). Plewa
et al. (2008) compared the level of cytotoxicity and genotoxicity of individual and
group DBPs with CHO cells and concluded that the order of cytotoxicity and
genotoxicity were iodo-DBPs > bromo-DBPs > chloro-DBPs (Figure 2.7). They also
reported that HNMs and HANs were more cytotoxic and genotoxic than HAAs and
THMs (Figure 2.7). For example, DCAA was reported 27 times more concentrated than
DCNM, but the toxicity of DCNM was 30.8 times the one of DCAA (Plewa et al.,
2004). Other chemicals to be of health concerns are the group of HKs, because 1,1-DCP
and 1,1,1-TCP have shown to exert carcinogenic and mutagenic effect in mice (Bull and
Robinson, 1986).
DBP chemical class
Iodo-DBPs
Bromo-DBPs
Chloro-DBPs
Halonitromethanes
Haloacetonitriles
Haloacetic acids
Genotoxicity
Cytotoxicity
Halomethanes
1,00E+02
1,00E+03
1,00E+04
1,00E+05
1,00E+06
CHO cell cytotoxicity or genetoxicity index values (log scale)
Figure 2.7 Cytotoxicity and genetoxicity indices for different classes of DBPs and for
chloro-, bromo- and iodo-DBPs (Plewa et al., 2008)
46
Despite their potential health effects, there is no UK or US regulatory limit for these
compounds, but the WHO has suggested guideline values of 20 µg/L for DCAN, 70
µg/L for DBAN and 10 µg/L for TCA (WHO, 2006).
2.4.2
Analytical measurement
Currently, there is one GC/ECD US EPA approved method: the US EPA Method 551.1
(1995a), which enables the measurement of 8 semi-volatile DBPs (DCAN, TCAN,
BCAN, DBAN, TCA, TCNM, 1,1-DCP and 1,1,1-TCP), the four THMs, 8 chlorinated
solvents and 17 pesticides/herbicides. The DBPs are extracted with LLE from water
samples using MtBE and analysed using GC/ECD. The advantages of this method are
the short time required for sample preparation (10 minutes) and the short run time to
process the 4 THMs and the 8 semi-volatile DBPs (less than 30 minutes). On the other
hand, the number of semi-volatile compounds analysed with the US EPA Method 551.1
is limited. Studies on emerging semi-volatile DBPs have therefore focused on
synthesised standards that were not previously commercially available and developed
new methods that incorporate approximately 50 compounds (Gonzalez et al., 2000;
Krasner et al., 2001; Weinberg et al., 2002). These methods assessed, include modified
versions of the LLE – GC/ECD US EPA Method 551.1 (1995a), a solid phase
extraction (SPE) – GC/MS to have MS confirmation of the semi-volatile DBPs (Krasner
et al., 2001). Weinberg et al. (2002) also used a purge-and-trap (P&T) – GC/MS for
certain volatile chemicals. Amongst these methods, the LLE has the best extraction
efficiency (Chinn et al., 2007). All these methods have been shown suitable to detect the
semi-volatile DBPs in drinking water. Nevertheless, before extraction, two preservative
parameters need to be controlled to avoid any DBP degradation: the quenching agent
and the pH (Chinn et al., 2007). Chinn et al. (2007) showed that many of the DBPs were
stable in ascorbic acid (quenching agent), but others, such as tribromoacetonitrile or
tribromonitromethane degraded over time in ascorbic acid, and therefore ammonium
chloride was recommended instead. Also, it has been shown that TCAN can undergo
base-catalysed hydrolysis at pH higher than 5.5 (Croué and Reckhow, 1989), hence,
waters with pH higher than 5.5 may present low amount of TCAN.
47
2.4.3
Typical levels in drinking water
A broad screen of the published works revealed that the concentration of most of the
compounds did not exceed 10 µg/L individually (Table 2.12), the exception being TCA
and DCA (Krasner et al. 2006). The HA group represents the third major class of
halogenated DBPs in weight basis after the THMs and the HAAs (Krasner et al., 2001).
Treatment applied played an important role on the formation of the DBPs. For example,
Bichsel and Von Gunten (2000) reported that i-THMs are favoured by chloramines.
Moreover, HNMs have been reported to be in higher concentration in waters preozonated (Hoigné and Bader, 1998; Richardson et al., 1999).
Table 2.12 Typical levels of semi-volatile DBPs found in drinking water
DBPs
Treatment applied
DCIM
BCIM
Chloramines
DCIM
BCIM
DBIM
Level observed
(µg/L)
0.30
0.20
0.50
Chlorine
Reference
Krasner et al. (2001)
0.60
0.60
Sum of DCIM,
BCIM, DBIM,
CDIM, BDIM
and TIM
Chlorine and ammonia
simultaneously
19.0
DCIM, BCIM,
DBIM
Chloramines
< 1.00
Cancho et al. (2000)
DCAN
Chlorine dioxide, chlorine and
chloramines
0.10 – 5.00
DBAN
Chlorine and chloramines
0.10 – 3.00
TCAN
Chlorine
0.10
Weinberg et al.
(2002)
Sum HANs
Chlorine
10.3 – 33.6
1,1-DCP
1,1,1-TCP
TCA
DCA
Ozone and chlorine
Not detected.
48
NDa – 5.00
13.0
Ozone and chloramines
0.30
3.00
Ozone and chloramines
12.0
2.00
Chlorine dioxide, chloramines
a
ND – 2.00
TCNM
Sum HNMs
Kim et al. (2002)
a
10.0
2.4.4
Parameters affecting semi-volatile DBP formation
Until now, few studies have focused on the impact of precursors on the semi-volatile
DBP formation. However, the first results stated that organic and inorganic precursors
can alter the formation of DBPs. Lee et al. (2007) investigated the impact of dissolved
organic nitrogen (DON) and concluded that DON can serve as a precursor material for
the nitrogen-containing DBPs such as DCAN and TCNM. Bromide and iodide are also
closely related to the formation of bromine and iodine-containing DBPs. Waters with
high level of bromide (≥ 500 µg/L) formed more brominated HANs (BCAN and
DBAN) than DCAN, and increasing the level of bromide decreased the formation of
chlorinated DBPs, such as 1,1-DCP (Peters et al., 1990; Yang et al., 2007). Recently,
Goslan et al. (2009) concluded that the concentration of i-THMs typically increases
with increasing iodine level.
Water treatment variables can also influence the formation of the semi-volatile DBPs.
Previously, the pH has been reported to be responsible of base-catalysed hydrolysis of
TCAN (Section 2.4.2). The choice of disinfectants also affects the level of DBPs
(Section 2.4.3). Recently, Yang et al. (2007) showed that DCAN and 1,1-DCP increased
with increasing reaction time; 90% of 1,1-DCP and over 60% of DCAN being formed
within seven hours and three days respectively, compared to their formation after seven
days. The same study also reported that DCAN and 1,1-DCP increased with higher
monochloramine doses and that the temperature did not alter the formation of DCAN
and 1,1-DCP.
2.5
CHAPTER SUMMARY

Several analytical methods have been reported in the literature for the
measurement of HAAs. Nevertheless, few of them are suitable for the
quantification of low µg/L level, expected in drinking water.

Regarding toxicity of DBPs, work is ongoing to assess the risks to humans. It is
thought that bromine-, iodine-containing DBPs pose more risk to health. Semivolatile DBPs, such as HNMs, or HANs, are more cytotoxic and genetoxic than
HAAs and THMs.
49

Typical levels of HAAs in drinking water ranged from 0 to 244 µg/L (total
HAA), and semi-volatile DBPs (i-THMs, HANs, HNMs, HAs, HKs) between 0
and 13 µg/L (individual species), with the highest concentration for HA, which
is therefore the third prevalent group of DBPs after the HAA and the THM.

Some studies reported that the hydrophilic NOM contributes substantially to the
formation of DBPs. However, other studies contradict this statement. DBP
precursors vary from one source to another and this explains why no specific
trend could be identified.

Strong and weak linear correlations have been observed between HAAs, THMs
and water characteristic parameters.

Formation of HAA will be maximised with a high DOC concentration, a high
bromide concentration, a high SUVA value and a high temperature coupled with
a high chlorine dose for a long contact time. HAA and THM formation will be
maximise with a low and high pH respectively.

In general, bromide and iodide, naturally present in drinking water source,
enhance the formation of bromide and iodide-containing DBPs.

The use of monochloramine, as an alternative disinfectant, has been found to
minimise the formation of HAAs and THMs, provided the conditions are correct
(a high pH with a low Cl2:N ratio). The formation mechanism between
monochloramine and NOM to form DBPs remains uncertain.

Predictive models for the formation of HAAs under-estimate the observed level
of HAAs. Developing accurate model is a complicated and difficult task that
requires a considerably larger body of data and has to be water specific.
50
Chapter 3
ANALYTICAL METHODOLOGY FOR HAAs
3.1
INTRODUCTION
Disinfection is well known to be effective in preventing waterborne disease, but the
reaction between chemical disinfectants and naturally occurring organic matter can form
DBPs of potential health concern (Christman et al., 1983). A great deal is known about
THM, but much less on HAA, often the second most prevalent group of DBPs formed
during chlorination (Singer, 2002) (see Chapter 2). Nine species of HAAs are found in
chlorinated water: MCAA, MBAA, DCAA, BCAA, TCAA, DBAA, BDCAA, DBCAA
and TBAA, and, of these, five are regulated in the US. The five HAAs, MCAA,
MBAA,
DCAA,
DBAA
and
TCAA,
are
regulated
under
the
Stage
1
Disinfectants/Disinfection Byproducts Rule (D/DBPR) with a maximum contaminant
level of 60 µg/L (US EPA, 1998). In the UK, no regulation has been set and only one
study has reported levels of HAAs in UK waters (Malliarou et al., 2005). They found
concentrations that were significantly above the US EPA regulatory limits.
To monitor HAAs in drinking water, several analytical methods can be used, including
GC/ECD, GC/MS, HPLC, CE and IC, where the GC methods are the most widely
applied. Currently, there are four GC/ECD US EPA approved methods: the US EPA
Method 552.1 (1992), US EPA Method 552.2 (1995), US EPA Method 552.3 (2003)
and the Standard Method 6251 (APHA, 1998). In these methods, HAAs are extracted
from water samples using either MtBE or anion exchange resins, and then converted to
their methyl esters using diazomethane or acidic methanol. Several shortcomings of the
three oldest methods (US EPA 1992, 1995; APHA, 1998) have been reported (Xie,
2001). Reduced susceptibility to chromatographic interferences and shorter run times
were observed with GC/MS (Xie, 2001), but the MS detection significantly increases
the price of analysis (Liu and Mou, 2003). In general, GC methods involve labourintensive extraction procedures and the use of toxic derivatisation reagents, but they are
typically reliable and accurate with detection limits for the nine HAAs in the low µg/L
51
Chapter 3 – Analytical methodology for HAAs
range. HPLC, which is a method faster than GC method, can also be used for the
determination of HAAs, but its detection limits are above 30 µg/L (Carrero and Rusling,
1999). CE is a suitable alternative to the chromatographic methods and has the
advantages of high separation efficiency, short analysis time and minimal sample
preparation (Xie and Romano, 1997; Xie et al., 2000). However, the drawback of CE is
its current detection limit that allows detection only in the mg/L range (Kim et al., 2001;
McRae et al., 2004). The IC method has been investigated by Nair et al. (1994) and Liu
and Mou (2003) and again the method is quick with limited sample preparation. Liu and
Mou (2003) reported detection limits for the nine HAAs between 0.37 and 31.64 µg/L.
More details on analytical measurement are described in Chapter 2, Section 2.3.2.
For this project, a convenient and sensitive method for the direct determination of
HAAs was required. Given the advantages and drawbacks for each analytical method,
the choice was not straightforward. The aims of this chapter are to compare four
analytical methods and to validate the most suitable one for the measurement of HAAs.
Here, a suitable method is regarded as one that has a minimum reporting level (MRL) of
0.5 µg/L (lowest continuing calibration standard) and therefore a detection limit of 0.17
µg/L for each HAA (US EPA, 2003). IC was investigated first as it offered a fast and
cheap method, and subsequently three GC methods were evaluated. The GC methods
were (1) GC coupled with MS, (2) a two dimensional GC coupled with time of flight
mass spectrometry (TOFMS) and (3) GC coupled with ECD. The most suitable method
was then validated according to the US EPA methods (1995, 2003).
3.2
METHOD COMPARISON FOR HAA ANALYSIS
The purpose of this section was to compare four analytical methods, to determine the
most suitable one for the quantification of HAAs.
3.2.1
Materials and Methods: IC
3.2.1.1 Reagents and glassware
Standards were individually available from Sigma-Aldrich Ltd (UK). The standards
used were MCAA at 99% purity, MBAA at 99+%, DCAA at 99+%, BCAA at 97%,
52
TCAA ACS reagent at 99+%, DBAA at 97%, BDCAA, neat, at 99+%, DBCAA, neat,
at 99+%, and TBAA at 99%. Reagent water used in the preparation of calibration
standards and blanks was prepared using PURELAB Ultra Genetic 18.2 MΩ-cm pure
water (ELGA LabWater, UK).
All glassware was meticulously washed with deionised water, soaked in a 5% nitric acid
solution for 12 hours, rinsed with deionised water three times and heated in a standard
dryer until thoroughly dry. Because the accuracy of conical flasks can be affected by the
standard dryer, they were placed on a drying rack until thoroughly dry.
3.2.1.2 Preparation of standards
The stock standard solutions were prepared in pure water. Stock solution 1 was
prepared at a concentration of approximately 10 mg/L for each compound. A secondary
standard solution at a concentration of 100 µg/L was prepared for each compound
individually. Calibration standard concentrations were 5, 25, 50 and 100 µg/L. The
concentrations 5, 25 and 50 µg/L were achieved by mixing and diluting the secondary
stock solution. The concentration 100 µg/L was achieved by mixing and diluting the
stock solution 1. Calibration standards were made freshly. Stock solutions were kept
refrigerated and were discarded after one month.
3.2.1.3 Method description
Before commencing the analysis, all 5 mL standards/samples were treated with two
cartridges. The first contained silver to remove chloride and the second was used to
remove any residual dissolved silver (OnGuard II Ag and OnGuard II H, Dionex, UK).
This method was first reported by Liu and Mou (2003). No chlorine was added to the
standard, but cartridges were used to make sure that the standards were treated the same
as the samples.
The analysis was performed at room temperature with all samples injected in duplicate.
The following HAAs were determined: MCAA, MBAA, DCAA, BCAA, TCAA,
DBAA, BDCAA, DBCAA and TBAA.
53
3.2.1.4 Instrumentation conditions
Analysis of the nine HAAs was carried out with an IC system (Dionex, DX500 series,
UK). This consisted of an Ion Pac AG9HC guard column (4 x 50 mm), an IonPac
AS9HC separation column (4 x 250 mm), an anion electrolytic suppressor (Atlas, 4
mm, UK) in auto-suppression recycle mode, an SC 20 suppressor controller, an
electrochemical detector (ED 40) in conductivity mode and a 250 µL sample loop (all
supplied by Dionex, UK). The eluent was 11.5 mM sodium carbonate (Na2CO3)
anhydrous with which all the HAAs could be well separated and detected. The
suppressor current was set to 85 mA and the eluent flow rate was kept at 1.0 ml/min.
This method was first reported by Liu and Mou (2003).
3.2.2
Materials and Methods: GC
The GC methods were (1) GC coupled with MS, (2) a two dimensional GC coupled
with TOFMS and (3) GC coupled with ECD. The procedure for standard/sample
preparation was identical for the three GC methods. The HAAs measured with GC
methods were MCAA, MBAA, DCAA, BCAA, TCAA and DBAA and are referred as
HAA6. At the time of this study, only these six HAAs were commercially available at
the same concentration in a certified mixture.
The solvent use for the extraction work was MtBE, HPLC grade, (Fisher Scientific,
UK). The internal standard was 99+% pure 1,2,3-trichloropropane (Fisher Scientific,
UK). The acidic methanol solution was prepared with concentrated sulphuric acid
(H2SO4), density of 1.83 g/mL and >95% pure, and reagent grade methanol (Fisher
Scientific, UK). Sodium sulphate (Na2SO4) (Fisher Scientific, UK) was 99% extra pure
anhydrous and was baked overnight at 100°C before the extraction. The standards were
available as EPA 552 halogenated acetic acids mix in a 1 mL ampoule from SigmaAldrich Ltd (UK). Purities for the HAA6 ranged between 97.4 and 99.9% and were
supplied at 2000 µg/mL each. Reagent water used in the preparation of calibration
54
standards and blanks was prepared from PURELAB Ultra Genetic 18.2 MΩ-cm pure
water (ELGA LabWater, UK).
Stock solution 1 was prepared in MtBE at a concentration of approximately 20 mg/L for
each compound by accurately transferring 50 µL of the EPA halogenated acetic acids
mix to a 5 mL volumetric flask. Calibration standard solutions were prepared by
diluting stock 1 in 100 mL volumetric flask with pure water. Stock 1 and calibration
standards were prepared freshly and discarded after use.
HAAs are highly water-soluble DBPs that exist as ions at ambient pH. They must be
converted to their volatile methyl ester form to be analysed with GC (Singer et al.,
2002). The derivatisation method used here was reported by Tung et al. (2006) and is a
modified version of US EPA Method 552.2 (1995) (Figure 3.1).
A 30 mL volume of sample was adjusted to a pH of 0.5 or less and extracted with 3 mL
of MtBE containing an internal standard. The protonated HAAs that have been
partitioned into the organic phase were then converted to their methyl esters by the
addition of acidic methanol followed by heating at 50ºC for 2 hours. The solvent phase
containing the methylated HAAs was separated from the acidic methanol by addition of
4 mL of a 10% aqueous solution of sodium sulphate. The aqueous phase was discarded.
The solvent phase was removed for analysis with (1) GC/MS, (2) GCxGC/TOFMS and
(3) GC/ECD.
55
Transfer 30 mL of water sample into a 60 mL glass vial with PTFE lined cap
Sample Extraction
Acidification with 1.5 mL concentrated sulphuric acid
Add exactly 3 mL of MtBE with internal standard (IS)
IS is 1 µg/mL of 1,2,3-trichloropropane
Add approximately 12 g muffled sodium sulphate
Shake for 3 minutes and allow layer separation for 5 minutes
Transfer approximately 1 mL of organic upper layer to Hach test tube
Add 1 mL of 10% sulphuric acid / methanol solution
Sample Methylation
Place tube in heating block at 50°C ± 2°C and heat for 2 hours
Add approximately 1 mL MtBE without IS
Add 3 mL of 10% sodium sulphate solution and vortex for 30 seconds
Remove aqueous phase as much as possible
Add 1 mL of 10% sodium sulphate solution and vortex for 30 seconds
Transfer 1 mL of upper layer to GC vials
Analysis by GC/MS
Analysis by GCxGC/TOFMS
Analysis by GC/ECD
Figure 3.1 Sample preparation procedure (Tung et al., 2006) – Adapted from US EPA
Method 552.2 (1995)
Optimisation of the US EPA Method 552.2 (1995)
Under the US EPA Method 552.2 methylation conditions, a complete methylation was
observed for MXAAs and DXAAs, but not for the four TCAAs (Xie et al., 2002). It was
concluded by the same study that improvement of methylation efficiency could be
56
achieved by working with an MtBE:acidic methanol ratio of 1:1, which is applied here.
Also, in the revised method, sodium sulphate solution (Na2SO4) was used to replace
sodium bicarbonate (NaHCO3) recommended in the US EPA Method 552.2 (1995).
Indeed it was reported that the release of carbon dioxide (CO2) from sodium bicarbonate
may result in a loss of HAA methyl esters (La Guardia, 1996).
GC/MS
HAA standards were run on a GC Perkin Elmer AutoSystem XL coupled with a
TurboMass Gold MS using the method reported by Xie (2001). This method involved a
gas flow rate of 1.0 mL/min. The oven temperature was set at 35°C and raised gradually
to 185°C. The injector temperature was 200°C, the ion source temperature was 230°C
and the electron energy was 70 eV.
To confirm the results found here, samples were also run on a GC Agilent 5973
interfaced with quadrupole MS. An injection volume of 1 µl of the HAA calibration
standard was introduced splitless into a BPX 5 GC column (SGE, UK; 30 m  0.25 mm
 0.25 m). The initial GC oven temperature was set at 55ºC and held for 2 minutes.
The temperature was then raised at a rate of 5ºC/min to 220 ºC. The GC injector
temperature was maintained isothermally at 200ºC. A constant flow rate of 99.999 %
pure inert helium gas was held at 1 ml/min. The MS parameters used were: transfer line
temperature 280ºC, manifold temperature and source temperature 230ºC, and the
electron energy was 70 eV.
Total ion chromatograms were obtained (m/z 35 – 300) and the following two ions, m/z
59 and 75 were selected for analysis. The m/z 59 ion is the base peak for the HAA
methyl esters and m/z 75 is the base peak for the IS. The MassTransit software was used
for data conversion and an Agilent ChemStation G1701DA was used for data handling
and processing.
57
GCxGC/TOFMS
HAA samples were run in parallel by comprehensive two dimensional GC/MS utilising
a Leco Pegasus VI GCGC/TOFMS (Figure 3.2). GCGC separation was performed
using an Agilent 6890 GC with a Leco GCGC modulator fitted coupled to a Pegasus
IV TOFMS (LECO Corporation, Michigan, US).
1
2
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
Sample
Inlet
First-dimension column
Modulator
Secondary oven
Second-dimension column
Transfer line
Ion source
Flight tube
Time-array detector
Instrument control/data
processing computer
4
7
8
6
5
3
10
9
11
Figure 3.2 Simplified diagram of GCxGC/TOFMS Instrument
The GC injector was operated in splitless mode with a column flow rate of 1.0 mL/min
and held at 200°C. GCGC separation utilised a non-polar column and a polar column:
a BPX5 (SGE, UK; 30 m  0.25 mm  0.25 m) and a BPX50 (SGE, UK; 1.8 m  0.1
mm  0.1 m) respectively. The GC oven temperature was held for 1 minute at 35°C
and ramped to 220°C at a rate of 5°C/min and then held for 1 minute, the second
column was ramped at 30°C above the first column. Modulation time was 4 seconds.
Mass spectra were acquired in electron ionisation mode from 33 to 400 amu with an
acquisition rate of 133 spectra per second.
58
GC/ECD
HAAs were also measured on a GC with an ECD (Agilent 6890). A volume of 1 µL
was injected with the injector set at 200°C with a 5:1 split ratio. The separation was
performed using a BPX5 column (SGE, UK; 30 m × 0.25 mm × 0.25 µm) with a helium
carrier gas at a column flow rate of 1.1 ml/min. The initial oven temperature was 35°C
followed by a 5°C per minute temperature ramp to 220°C and held for 1 minute. The
detector temperature was 230°C and the rate of data collection 20 Hz.
3.2.3
Results and Discussion
3.2.3.1 IC method
HAA standards were first run individually with IC to determine their retention time. A
typical IC chromatogram for a 10 µg/L mixture of HAA9 is shown (Figure 3.3). As
shown in Table 3.1, all the coefficients of determination of linear regression were above
0.9644, which expressed a linear calibration curve for each of the nine HAAs. This
compares well with Liu and Mou (2003) (Table 3.1).
Table 3.1 Calibration correlation coefficients for HAA9 analysed by IC in this present
study and in a study by Liu and Mou (2003)
HAA
MCAA
MBAA
DCAA
BCAA
DBAA
TCAA
BDCAA
DBCAA
TBAA
Correlation coefficients (R2)
Present study
Liu and Mou (2003)
0.9997
0.9992
0.9985
0.9988
0.9991
0.9990
0.9997
0.9989
0.9943
0.9886
0.9916
0.9993
0.9888
0.9991
0.9890
0.9884
0.9644
0.9889
59
Figure 3.3 Chromatogram of HAA9 standards (10 µg/L each). Retention time (minutes):
MCAA = 5.547, MBAA = 5.960, DCAA = 9.767, BCAA = 11.367, DBAA = 12.103,
TCAA = 28.627, BDCAA = 33.590, DBCAA = 40.500, TBAA = 49.873
Limits of detection (LOD) were calculated as three times the signal-to-noise ratio for
HAA6 (MCAA, MBAA, DCAA, BCAA, TCAA, DBAA) and HAA3 (BDCAA,
DBCAA and TBAA) and were defined as the smallest concentration that can be
determined which can be expected to be distinguishable from the blank measurement
(Quevauviller, 2002). LOD was determined by spiking calibration standards into
upland, ground and lowland water. The detection limits of the analytes achieved with
the column and the suppressor ranged between 2.6 and 16.5 µg/L for HAA6 and
between 11.7 and 31.9 µg/L for HAA3 (Table 3.2). Liu and Mou (2003) reported greater
LOD for HAA6 between 0.37 and 2.16 µg/L and comparable LOD for HAA3 between
8.17 and 31.64 µg/L in deionised water. Nair et al. (1994) found LOD for five HAAs
(MCAA, MBAA, DCAA, TCAA and DBAA) above 8.0 µg/L, which is higher than the
values reported here. The limits of detection observed here for HAA6 are higher than
those reported by Liu and Mou (2003) and were believed to be a result of the matrix
impact. The criterion to accept a method here was LOD of 0.17 µg/L for a single HAA;
hence, LOD for HAA3 was poor as none of them could be detected below 10 µg/L.
60
Table 3.2 Detection limits for the analytes from this present study, Liu and Mou (2003)
and Nair et al. (1994)
Waters
Ground water
Upland water
Lowland water
Deionised water (Liu and Mou,
2003)
Deionised water (Nair et al.,
1994)
a
HAA6 (µg/L)
5.0 – 16.5
2.6 – 11.0
3.2 – 14.1
HAA3 (µg/L)
21.4 – 26.8
13.8 – 23.0
11.7 – 31.9
0.4 – 2.2
8.2 – 31.6
8.0 – 80.0a
NRb
Data available for 5 HAAs (MCAA, MBAA, DCAA, TCAA and DBAA); b Not reported.
The choice of an IonPac AS9HC column is explained by Liu and Mou (2003). They
reported that fluoride, naturally present in the water, MCAA, MBAA and chloride are
weakly retained anions, and if the retention time of fluoride, MCAA and MBAA are
close to that of chloride, the high response of chloride will seriously interfere with the
other anions. Hence in order to separate these four anions well, a high capacity column
on which the anions could be tightly bound was preferred (i.e. IonPac AS9HC column).
With the use of carbonate as eluent, the nine HAAs were well separated. The Atlas
anion electrolytic suppressor, used here, exchanged the cations from the eluent (aqueous
Na2CO3) for hydronium ions (H3O+) forming carbonic acid (H2CO3) of low
conductivity, and exchanged sample for fully protonated HAAs of high conductivity.
Hence better detection sensitivity has been observed with suppressed IC than with nonsuppressed IC (Jauhiainen et al., 1999).
The advantage of the IC was the simplicity and the speed of sample preparation (about
10 minutes), but GC methods were reported to be more sensitive and were believed to
be more suitable for the measurement of HAAs in UK treated water. Hence no further
work was undertaken with IC.
3.2.3.2 GC methods
GC/MS was used to analyse six HAA standards that had been derived to their methyl
esters. The GC/MS method was reported by Xie (2001), and although he was able to
resolve all the nine HAAs with GC/MS (Figure 3.4 A), the peaks, in this present study,
were not well resolved nor was the signal-to-noise ratio sufficient. The GC/MS was run
61
in the selective ion monitoring (m/z = 59) mode but this did little to improve the
resolution and the sensitivity. It was not possible to determine limits of detection for this
method. In order to confirm the findings, samples were run in parallel using another
GC/MS (Agilent 5973) (Figure 3.4 B). The results were comparable to the GC/MS
Perkin Elmer Turbomass.
Figure 3.4 Comparison of a 100 µg/L of each derived HAA standard chromatograms
from two similar GC/MS instruments by (A) Xie (2001) and (B) this present study
To investigate the lack of resolution and sensitivity further, samples were run using a
Leco Pegasus 4D GC×GC/TOFMS. This machine uses two GC columns to separate
analytes based on volatility as well as polarity. The derived HAA methyl ester peaks
could be observed as they had been separated from the interfering material (Figure 3.5
A+B), but the interfering material had a greater intensity than some of the derived
methyl esters and also eluted at retention times that overlapped with the derived HAA
methyl esters (Figure 3.5 A). The interfering peaks are thought to be incurred from the
derivatisation procedure, but to date, have not been identified.
62
HAA6
A
B
Figure 3.5 (A) Total ion chromatogram for derived HAA6 (equivalent to 100 µg/l of
each HAA) and (B) a partially reconstructed mass chromatogram (m/z = 59) of a
derived HAA6 standard at 100 µg/l of each HAA (Leco Pegasus 4D GC×GC/TOFMS)
Despite the effectiveness and the enhanced separation of the GCxGC/TOFMS, no
further work was undertaken with this instrument due to availability and the cost of
analysis.
HAA analysis was then considered with GC/ECD. A derived HAA6 standard at 100
µg/L of each HAA was run with a GC/ECD (Agilent 6890). The six HAAs gave good
responses and no interfering peaks were observed (Figure 3.6). The advantage of the
ECD is the selectivity and approved methods reported LOD for HAAs in the low µg/L
range (US EPA, 1995, 2003). Indeed, the ECD is particularly sensitive to halogens
because the detection is based on how much the halogens capture electrons produced by
a Beta particle (electron) emitter (Lovelock, 1958; Lovelock, 1974). The GC/ECD
method, however, requires labour-intensive and fastidious extraction (6 hours).
Nevertheless it was found to be the most appropriate analytical device for the analysis
of HAAs in UK treated water and the validation work is presented in Section 3.3.
63
Figure 3.6 Chromatogram for a diluted (equivalent to 100 µg/l of each) and derivatised
HAA6 standard with GC/ECD
3.2.4
Conclusion
Here four different methods were investigated for the determination of HAAs. With IC,
the detection limits for HAA6 were between 2.6 and 16.5 µg/L and between 11.7 and
31.9 µg/L for HAA3. LOD for HAA9 were well above the level required to accept the
method and IC was not deemed sensitive enough for the work here. Due to interfering
peaks overlapping HAAs, GC/MS was not suitable for the measurement of HAAs.
GCxGC/TOFMS is a very selective and effective tool, but expensive. Despite its
labour-intensive requirements, GC/ECD was selected as the best suitable analytical
device for the quantification of HAAs for the remaining work in this thesis.
3.3
GC/ECD METHOD VALIDATION
The validation of nine HAAs was determined in compliance with the US EPA Method
552.2 (1995) and 552.3 (2003). The nine HAAs are referred as HAA9.
64
3.3.1
Materials and Methods
The reagents were identical to those in Section 3.2.2.1, except for the EPA 552.2
haloacetic acids mix in a 1 mL ampoule from Sigma-Aldrich Ltd (UK). Purities for the
HAA9 ranged between 96.2 and 99.9% and were concentrated at 2000 µg/mL each. All
glassware was treated as in Section 3.2.2.1. Ammonium chloride (NH4Cl) (Fisher
Scientific, UK) at a concentration of 100 mg/L was used as a quench agent and has been
shown to not degrade any HAAs, and especially HAA3 (BDCAA, DBCAA and TBAA)
(Singer et al., 2002).
Standard preparation was explained in Section 3.2.2.2.
The method was explained in Section 3.2.2.3 (Figure 3.1). Here the solvent phase was
removed for analysis with GC/ECD only.
HAA standards and samples were run with a different Agilent 6890 GC/ECD than the
one used in Section 3.2.2.4. A volume of 1 µL was injected with the injector set at
200°C with a 10:1 split ratio. Separation was performed by a ZB-1ms column
(Phenomenex, UK; 30 m  0.25 mm  0.25 m) with a helium carrier gas at a column
flow rate of 0.9 mL/min. The initial oven temperature was 35°C and held for 8 minutes
followed by an 8°C per minute temperature ramp to 200°C and held for 1 minute. The
total run time was 29.63 minutes. The detector temperature was 270°C and the rate of
collection data was 20 Hz.
65
3.3.1.5 Method performance
The average retention time for each HAA was determined with seven replicates of a
derived 1.0 µg/L of each HAA and internal standard aqueous standard. The retention
time precision was evaluated by calculating the relative standard deviation (RSD) as
recommended by US EPA Method 552.3 (2003). The RSD of the sample replicate
analyses must be less than 20% and was calculated with the following equation:
RSD (%) = (
Standard deviation of array X
)  100 .
Mean of array X
Equation 3.1
Seven replicates of a derived 0.10 µg/L and 1.00 µg/L of each HAA aqueous standard
were extracted and analysed over a period of three days for determining LODs.
Ammonium chloride, which was the HAA preservative agent, was added to the
standards at the same concentration as that used to quench the samples. LOD was
calculated using the following equation:
LOD = St ( n1, 1
 0.99 )
Equation 3.2
where: t(n-1, 1-α = 0.99) = Students t value for the 99% confidence level with n-1 degrees of
freedom,
n = number of replicates, and
S = standard deviation of the replicate analyses.
Blanks were not subtracted when performing LOD calculations.
The analyte recovery was determined by analysing two treated water samples spiked
with three levels of HAAs. Fortifications were at 5, 25 and 75 µg/L. The recovery for
each analyte was calculated using the following equation:
R=
(D  E)
100%
F
where: D = measured concentration in the fortified sample,
E = measured concentration in the unfortified sample, and
F = fortified concentration.
66
Equation 3.3
3.3.1.6 Water samples
Treated water samples used to determine the analyte recovery were geographically
different. The lowland and the upland waters were collected from South East England
and northern England respectively. The treatment of the lowland water consisted of
ozonation, coagulation, GAC and chlorination, whereas the treatment of the upland
water was coagulation, sand filtration and chlorination. Both treated samples were
collected prior to chlorination at the water treatment works.
3.3.2
Results and Discussion: Evaluation of method performance
HAA analysis with GC/ECD is widely used and ECD has been proven to be very
selective and sensitive to HAAs. However quality controls were required prior to
analyte quantification. Chromatograms for nine HAAs and internal standard were
obtained by running a standard using the chromatographic conditions described in
Section 3.3.1.4. The retention times for all nine HAAs and internal standard are listed in
Table 3.3. The RSD calculated for the retention times reported here were typically low
and were similar to those reported by US EPA Method 552.3 (2003), demonstrating
good reproducibility of the method.
Table 3.3 Identification of the compounds and method performance
Peak
no.
1
2
3
4
5
6
7
8
9
IS
Compound
MCAA
MBAA
DCAA
BCAA
TCAA
DBAA
BDCAA
DBCAA
TBAA
ISe
Average
tra (min)
6.69
9.86
10.42
13.21
13.37
15.37
15.76
17.80
19.83
12.97
RSDb
(%)
0.104
0.030
0.019
0.012
0.011
0.006
0.009
0.010
0.067
0.010
R2
0.999
0.999
0.996
0.999
0.996
0.998
0.994
0.991
0.987
NAf
Fortification
level (µg/L)
1.00
0.10
1.00
0.10
1.00
0.10
0.10
0.10
0.10
NAf
Detection
Limitc (µg/L)
0.783
0.086
0.317
0.026
0.064
0.022
0.037
0.055
0.045
NAf
MRLd
(µg/L)
2.349
0.258
0.951
0.078
0.192
0.066
0.111
0.165
0.135
NAf
a
The average retention time corresponds to the average of seven injections; b Corresponds to the relative
standard deviation and must be less than 20% according to US EPA Method 552.3 (2003); c Fortified
waters were extracted and analysed over 3 days for seven replicates; d Corresponds to the minimum
reporting level and is the threshold expected for accurate quantification in an unknown sample. It has to
be at least three times the limit of detection; e 1,2,3-trichloropropane; f Not applicable.
67
A typical GC/ECD chromatogram for a 100 µg/L of HAA9 derived standard is shown
(Figure 3.7) and peaks were identified in Table 3.3. The highest peak was DBAA
followed by BDCAA, TCAA and other HAAs. Similar responses were published in the
US EPA Method 552.3 (2003). The response is proportional to the electrons attracted
toward the halogen, which induces a reduction of the current generated by the ECD.
Although the electronegativity of chlorine is greater than that of bromine (3.16 and 2.96
Pauling respectively), MBAA captured more electrons than MCAA. This is believed to
be related to the intermolecular forces present in the methyl ester form of the HAAs.
MCAA had the smallest ECD response and this is supported by the US EPA Methods
552.2 (1995) and 552.3 (2003). The chromatograms presented in the US EPA Method
552.3 (2003) reported different orders of elution for BCAA, TCAA, DBAA, BDCAA
and internal standard compared to the method reported here. The main reason is
explained by the polarity of the column used for analysis. Here the column was a ZB1ms and has a lower polarity than the columns DB-1701 and DB-5.625 used in the US
EPA Method 552.3 (2003). The time spent by each solute in the stationary phase was
therefore variable depending on the column polarity. Therefore the retention factor,
which is the ratio of the amount of time a solute spends in the stationary phase and
mobile phase, varied for each species between columns. Unknown peaks were observed,
but did not interfere with HAAs and internal standard.
68
Figure 3.7 GC/ECD chromatogram for a derived HAA9 standard (100 µg/L) (see Table
3.3 for analyte identification)
Calibration was performed by extracted procedural standards, i.e. fortified pure water,
by the procedure in Section 3.2.2.3. The US EPA Method 552.2 (1995) recommended,
five calibration standards, here six calibration points were plotted and were 0.5, 5, 25,
50, 75 and 100 µg/L. The calibration curves were generated by plotting the area ratios
(AHAA/AIS) against the concentration CHAA of the six calibration standards (Figure 3.8)
where: AHAA = the peak area of the HAA
AIS = the peak area of the internal standard
CHAA = the concentration of the HAA.
A linear calibration curve was obtained for each of the nine HAAs and no HAA was
detected in pure water free of HAA spikes. The coefficients of determination of linear
regression are reported in Table 3.3 and were all above 0.987.
69
20
MCAA
MBAA
DCAA
BCAA
TCAA
DBAA
BDCAA
DBCAA
TBAA
18
Area HAA/Area IS
16
14
12
10
8
6
4
2
0
0
20
40
60
80
100
Figure 3.8 Calibration curves for nine HAAs – Concentration: 0.5, 5, 25, 50, 75 and 100
µg/L
The LOD (Table 3.3) for MCAA was 0.783 µg/L, which is higher than the values 0.273
µg/L and 0.17 µg/L reported by the US EPA Method 552.2 (1995) and US EPA Method
552.3 (2003) respectively. The variation of MCAA LOD is caused by the varying
GC/ECD sensitivity between studies. LOD for DCAA was 0.317 µg/L and is
comparable with the US EPA Method 552.2 (1995), and similar to that reported by
Malliarou et al. (2005) who found DCAA LOD at 0.8 µg/L using the US EPA Method
552.2 (1995). LODs for the remaining HAAs were greater than these reported by the
US EPA Method 552.2 (1995). The improved LOD is mostly due to the optimisation of
the experimental method. The total LOD for the nine HAAs was 1.435 µg/L. For
comparison, the LOD standards were sent to a collaborative laboratory (Open
University, Milton Keynes, UK), where they found a LOD of 1.494 µg/L for the nine
HAAs, which is very comparable. Malliarou et al. (2005) reported an LOD for 6 HAAs
(MBAA, DCAA, BCAA, TCAA, DBAA and BDCAA) of 4.9 µg/L. In relation to LOD,
the minimum reporting level (MRL) for each HAA was calculated (Table 3.3). The
MRL, which is at least three times the LOD, corresponds to the threshold expected for
accurate quantification in an unknown sample. Except for MCAA, the remaining HAAs
70
could be well quantified below 1 µg/l. It was decided to accept MRL below 0.5 µg/L
(e.g. BCAA at 0.078 µg/L) considering that the interpolation of the calibration curves
were good and there is a high likelihood that concentrations below the lowest
calibration standards are accurately quantified.
Analyte recoveries exhibited bias in both water samples (Table 3.4). The guideline
provided by the US EPA Method 552.3 (2003) recommends that recoveries should
range between 70 and 130%, except for the low level of fortification, where 50 to 150%
is acceptable. BCAA and DBAA were found to be biased in both waters. This is mostly
caused by matrix interferences, such as contaminants that are co-extracted from the
sample (US EPA, 2003). DCAA was also found to be biased in the upland water. This
may be due to the significant background of DCAA present in the unfortified matrix
(US EPA, 2003). The remaining HAAs were well recovered and RSD were acceptable
for all HAAs at all levels. Because some recoveries of DCAA, BCAA and DBAA fell
outside the designated range, concentrations of these analytes in real water would have
to be corrected to obtain meaningful values.
71
Table 3.4 Spike recovery in lowland and upland water fortified with HAAs
Compounds
MCAA
MBAA
DCAA
BCAA
TCAA
DBAA
BDCAA
DBCAA
TBAA
Compounds
MCAA
MBAA
DCAA
BCAA
TCAA
DBAA
BDCAA
DBCAA
TBAA
LOWLAND WATER SAMPLE
Fortification at 5 µg/L
Mean
Mean
a
RSD (%)
RSDa (%)
recovery
recovery
(n = 3)
(n = 3)
(%)
(%)
132
9.3
120
12.0
121
5.5
106
5.1
150
2.8
130
1.1
143
1.9
1.3
138 b
111
5.2
113
5.7
143
3.0
2.1
150 b
90
7.9
103
9.3
84
11.0
90
11.2
76
16.4
71
13.1
UPLAND WATER SAMPLE
Mean
Mean
RSDa (%)
RSDa (%)
recovery
recovery
(n = 3)
(n = 3)
(%)
(%)
122
3.6
119
5.2
132
3.5
117
2.2
b
b
3.7
1.1
153
131
145
1.4
1.1
136 b
119
3.2
115
3.1
146
2.3
1.5
142 b
100
6.0
103
5.1
96
7.4
93
6.7
88
10.3
81
8.5
Mean
RSDa (%)
recovery
(n = 3)
(%)
112
1.1
102
1.5
123
1.4
1.8
136 b
115
6.1
2.9
148 b
111
8.7
98
11.5
77
14.1
Mean
RSDa (%)
recovery
(n = 3)
(%)
117
1.9
109
1.1
121
1.4
1.1
133 b
117
2.1
1.5
141 b
114
4.8
103
6.5
86
9.6
a
Corresponds to the relative standard deviation and must be less than 20% according to US EPA Method
552.3 (2003); b Corresponds to biased values.
3.3.3
Conclusion
The US EPA Method 552.2 (1995), adapted by Tung et al. (2006) was validated here
for the quantification of nine HAAs. The method was found to be reproducible and
precise with some bias. Linear calibration curves were reported for each HAA. MCAA
LOD was high, and this was a result of the GC/ECD sensitivity. The LODs for the
remaining HAAs were comparable or greater than the guidelines and the total LOD for
the nine HAAs was 1.435 µg/L. Spike recovery calculations showed biased analytes.
DCAA, BCAA and DBAA recoveries were believed to be biased due to the matrix
effect and concentrations of these found in real samples should be corrected to give
realistic values.
72
3.4
CHAPTER SUMMARY
The results presented here showed that GC/ECD was the most suitable analytical device
for the measurement of nine HAAs during this thesis (Table 3.5).
Table 3.5 Summary of analytical method conditions and methods performance
Measuring
device
IC
Sample
preparation
Minimal
preparation
Cartridges to
remove chlorine
and silver
Analysis
time
Suitable for
present
study
LOD
(µg/L)
References
2.6 – 31.9
Present study
0.4 – 31.6
Liu and Mou
(2003)
< 50 mins
0.04 – 0.83
Xie (2001)
No
NRa
NRa
Present study
No
0.022 –
0.783
Present study
0.820 –
0.066
US EPA Method
552.2 (1995)
0.012 –
0.170
US EPA Method
552.3 (2003)
< 55 mins
No
~ 10 minutes
GC/MS
GCxGC/
TOFMS
GC/ECD
Extensive
preparation
Derivatisation
(extraction +
methylation) US
EPA methods
< 30 mins
~ 6 hours
a
Yes
Not reported.
Considering the good calibration curves, it was accepted MRL below 0.5 µg/L and
therefore, LOD below 0.17 µg/L for each HAA. The US EPA Method 522.2 (1995),
adapted by Tung et al. (2006), was validated according to US EPA Methods 552.2
(1995) and 552.3 (2003). Investigation of the reproducibility, precision and bias
demonstrated the reliability of the method.
73
74
Chapter 4
ANALYTICAL METHODOLOGY FOR
SEMI-VOLATILE DBPS
4.1
INTRODUCTION
Disinfected water has been shown to contain HAAs and THMs, but there is evidence to
suggest that significantly more than 600 chemical DBPs may also exist (Richardson,
1998). There is considerable uncertainty over the identity and levels of DBPs that
people are exposed to from drinking water. While most studies have focused on THMs
and HAAs, there is a need for comprehensive quantitative occurrence and toxicity data
to determine whether other DBPs pose an adverse health risk (Richardson et al., 2003).
To address this issue, US nationwide DBP occurrence studies have been undertaken
(McGuire et al., 2002; Weinberg et al., 2002). Approximately 50 DBPs, identified in US
waters, were rated high priority, because they showed an elevated level of cyto- and
genotoxicity (Krasner et al., 2001; Plewa et al., 2004). These DBPs include iodinatedTHMs, HANs, HAs, HKs and HNMs and are also referred to as semi-volatile DBPs,
due to their Henry’s law constant (Krasner et al., 2001). Because they have been
discovered more recently than THMs and HAAs and some of these semi-volatile DBPs,
such as the HANs, are unstable in aqueous solution (Peters et al., 1990), there has been
no regulation promulgated. Nevertheless the WHO has recommended guidelines for two
HANs (DCAN at 20 µg/L and DBAN at 70 µg/L) and for one HA (TCA at 10 µg/L)
(WHO, 2006). US studies reported levels of DCAN and DBAN below the WHO
recommendations (5 and 4 µg/L respectively), whereas TCA reached a maximum of 16
µg/L, which is beyond the WHO guidance (Krasner et al., 2001; Weinberg et al., 2002).
An occurrence survey in Canada detected TCA at a level up to 22.5 µg/L, which again
exceeds the WHO guidance (Williams et al., 1997). Peters et al. (1990) found levels of
less than 1 µg/L for DCAN and DBAN, which are below the US surveys and the WHO
recommendations. The other semi-volatile DBPs, such as the i-THMs, HNMs or HKs,
have also been detected in treated water at levels up to 19, 10 and 9 µg/L respectively
75
Chapter 4 – Analytical methodology for semi-volatile DBPs
(Krasner et al., 2006). All these semi-volatile DBPs of health and regulatory concern are
being identified in disinfected treated water from all chemical disinfection processes
(Krasner et al., 2006) and are therefore believed to be present in UK treated water. To
the knowledge of the author, no study has been published reporting levels of i-THMs,
HANs, HAs, HKs and HNMs in UK treated water.
In order to determine and to quantify these semi-volatile DBPs, reliable and
reproducible methods are necessary. Currently, there is one GC/ECD US EPA approved
method: the US EPA Method 551.1 (1995a). This method enables the measurement of 8
semi-volatile DBPs (DCAN, TCAN, BCAN, DBAN, TCA, TCNM, 1,1-DCP and 1,1,1TCP), four THMs (TCM, BDCM, DBCM and TBM), 8 chlorinated solvents and 17
pesticides/herbicides. In this method, DBPs are extracted from water samples using
MtBE and analysed using GC/ECD. The advantages of this method are the short time
required for sample preparation (10 minutes) and the short run time to process the 4
THMs and the 8 semi-volatile DBPs (less than 30 minutes). On the other hand, the
number of semi-volatile compounds analysed with the US EPA Method 551.1 is
limited. Studies on emerging semi-volatile DBPs have focused on synthesised standards
that were not previously commercially available and developed new methods that
incorporate approximately 50 compounds (Gonzalez et al., 2000; Krasner et al., 2001;
Weinberg et al., 2002). These methods assessed, include modified versions of the LLE –
GC/ECD US EPA Method 551.1 (1995a) and a SPE – GC/MS to have MS confirmation
of the semi-volatile DBPs (Krasner et al., 2001). Quality control demonstrated good
reliability of these methods, but SPE methods are more expensive and take more time
than LLE. More recently a study by Chinn et al. (2007) reported an automated SPE
method combined with GC/ECD and GC/MS for the analysis of 35 DBPs instead of
LLE, but reported LLE as having better extraction recoveries compared to SPE.
An insight into the semi-volatile DBP levels in UK treated water was of interest and
hence a reliable method for their quantification was required. After comparison of levels
of semi-volatile DBPs found in a previous study by Weinberg et al. (2002), 16 DBPs
were targeted. The targeted compounds encompassed 12 semi-volatile DBPs: two iTHMs, four HANs, two HAs, two HKs and two HNMs. In addition the four regulated
THMs were added to the method. Compared to the US EPA Method 551.1 (1995a), the
additional compounds were DCA, DBNM, DCIM and BCIM. By comparison with the
76
surveys from the Canada, Netherlands and US, previously cited (Peters et al., 1990;
Williams et al., 1997; Krasner et al., 2001; Weinberg et al., 2002), it is believed that the
analytical method here has to be reliable for low range of µg/L concentrations.
Therefore, criteria set for the method are based on accuracy, precision, working
range/linearity, selectivity and limit of detection (Fischbacher, 2000). To determine if
the method was suitable for the work undertaken here, the MRL chosen was 0.5 µg/L
(the lowest standard), which induced a LOD of 0.17 µg/L (Krasner et al., 2001).
The objectives of this chapter were (1) to identify qualitatively the presence of semivolatile DBPs in UK treated water using SPE – GC/ECD, and (2) to develop a LLE –
GC/ECD analytical method for the quantification of the targeted compounds. The
primary objective of this chapter was undertaken under the supervision of Dr. Howard
Weinberg at the Environmental Sciences and Engineering department laboratory of
University of North Carolina (UNC, Chapel Hill, US). Work continued in the UK at
Cranfield University to complete the secondary objective of this chapter.
4.2
QUALITATIVE IDENTIFICATION OF THE SEMI-VOLATILE DBPs
This is the first attempt to qualify many of these semi-volatile DBPs in UK treated
waters. A selection of 12 semi-volatile DBPs plus four THMs (Table 4.1) was made
after determining the most prevalent species in 12 drinking US waters (Weinberg et al.,
2002). SPE was used to concentrate up the DBPs to determine their presence in treated
waters, offering an alternative extraction means to conventional LLE (Weinberg et al.,
2002). With the SPE method used here, 100 mL of chlorinated water was concentrated
to 2 mL providing a sufficient concentration factor of 50 fold to achieve low µg/L
detection.
77
4.2.1
The solvent used for SPE was MtBE (Ultra-Resi Analysed grade) from J.T. Baker
(Phillipsburg, NJ, US). HPLC-grade methanol (Fisher Scientific, Pittsburgh, PA, US)
was used to condition the cartridges. All the standards used here were commercially
available (Table 4.1). The four THMs were available as Trihalomethanes Calibration
Mix in a 1 mL ampoule. Four HANs, two HKs and one HNM were available as EPA
551B Halogenated Volatiles Mix in a 1 mL ampoule. All the other standards were
available as individual species. The stock solutions were prepared in acetonitrile
(CH3CN) (Fisher Scientific, Pittsburgh, PA, US) and MtBE.
Table 4.1 Compound purity and suppliers
Compounds
Trihalomethanes
TCM
BDCM
DBCM
TBM
Haloacetonitriles
DCAN
TCAN
BCAN
DBAN
Haloketones
1,1-DCP
1,1,1-TCP
Halonitromethanes
TCNM (also called
chloropicrin)
DBNM
Haloacetaldehydes
DCA
TCA (also called chloral
hydrate)
Iodo-Trihalomethanes
DCIM
BCIM
a
Conc.a
(µg/mL)
Purity
(%)
Suppliers
2000
97.5
98.9
96.9
99.9
Sigma-Aldrich Ltd (St.
Louis, MO, US)
2000
99.9
99.9
99.1
95.2
Acetone
2000
96.9
97.8
Acetone
2000
99.9
NAb
Neat
> 90
NAb
Neat
> 90
Acetonitrile
1000
99.9
NAb
NAb
Neat
Neat
> 90
> 90
Stock solvent
Methanol
Acetone
Louis, MO, US)
Helix Biotech (Canada)
TCI America (Portland,
OR, US)
Louis, MO, US)
Helix Biotech (Canada)
Concentration; b Not applicable.
Reagent water used in the preparation of calibration standards and blanks was prepared
in the UNC laboratory using a water purification system (Pure Water Solutions,
78
Hillsborough, NC, US). The system filters chloraminated tap water to 1 µm, removes
residual disinfectants, reduces TOC to less than 0.2 mg/L with activated carbon, and
reduces conductivity to 18 MΩ-cm with mixed bed ion-exchange resins.
All non-volumetric glassware was detergent washed in a dishwasher, and then soaked in
a 10% nitric acid bath. The glassware was removed and rinsed three times with
deionised water and then placed in a 110°C oven over night or until thoroughly dry
(about 4 hours). Volumetric glassware (pipettes, volumetric flasks, and graduated
cylinders) was manually detergent-washed with tap water, rinsed three times with
deionised water, then methanol, and allowed to dry on clean tissues (Kimwipes) in a
dust-free environment. Caps and septa were soaked in a clean beaker filled with a
mixture of deionised water and detergent for an hour or more, rinsed with deionised
water, methanol, and then dried on a clean tissue.
The initial stock standard solutions for DBNM, DCA, DCIM and BCIM were first
prepared in 10 mL acetonitrile from the neat compounds. Acetonitrile was
recommended by Chinn et al. (2007) since the compounds have been observed to be
stable in this solvent. The purity of these four compounds was below 95%; hence a
correction factor was applied for the adjustment of the concentrations. For example
23.80 mg of DBNM was weighed, thus its accurate mass was actually 23.80 x 90% =
21.42 mg. The same calculations were applied for DCA, DCIM and BCIM. Therefore
the initial stock solution concentrations for DBNM was 2142 µg/mL, 1760 µg/mL for
DCA, 2304 µg/mL for DCIM and 2448 µg/mL for BCIM.
Stock solution 1 was prepared at a concentration of approximately 40 mg/L for each
compound by transferring 100 µL of the THMs Calibration Mix pure material, 100 µL
of the 551B Halogenated Volatiles Mix pure material, 200 µL of TCA pure material,
93.4 µL of DBNM initial stock solution, 113.5 µL of DCA initial stock solution, 87 µL
of DCIM initial stock solution and 82 µL of BDIM initial stock solution into seven 5
mL volumetric flasks containing acetonitrile. Secondary standard solutions at 100 µg/L
were prepared by diluting 25 µL of the stock solutions 1 in 10 mL volumetric flasks
containing MtBE.
79
Initial standard solution and stock solution 1 were kept at -20ºC and were discarded
after two weeks. Secondary standard solutions were made freshly on the day of analysis.
4.2.1.3 Water sample preparation
Treated water samples, used to determine the presence of the 16 DBPs, were
geographically different. A lowland and an upland water were collected from South
East England and northern England respectively. The treatment of the lowland water
consisted of ozonation, coagulation, GAC and chlorination, whereas the treatment of the
upland water was coagulation, sand filtration and chlorination. Both treated samples
were collected prior to chlorination at the WTWs.
DBP-FP was carried out using an adaptation of procedure 5710 A in “Standard Methods
for the Examination of Water and Wastewater” (APHA, 1992) and is described in more
detail in Chapter 5, Section 5.2.3. An appropriate volume of chlorine dosing solution
was added to a 100 mL bottle with PTFE-lined screw cap. Then the bottles were filled
completely with water sample and were stored at 20 ºC for 3 days. No buffer was added
to the samples. The final concentration of chlorine solution was 10 mg/L and was
sufficient to form DBPs for the qualification work (Chapter 5, section 5.2.3).
SPE was performed using a commercially available 12-port Visiprep vacuum manifold
and
1/8-inch
Teflon
tubing
with
weighted
stainless
steel
ends
(Supelco
Chromatography, Bellafonte, PA, US). The SPE cartridge contained styrene divinyl
benzene (DVB) polymeric beads (Strata; SDB-L; 500 mg, 3 mL, Phenomenex, US), and
has been shown to be effective for sorbing all the DBPs investigated here (Chinn et al.,
2007). Samples (100 mL) were placed in 125 mL conical flasks with the top of each
sample vial wrapped with aluminum foil and the edges of the Teflon tubing sealed with
a rubber band. The method is adapted from Weinberg et al. (2002) and Chinn et al.
(2007) and is described in Figure 4.1.
80
Four 3 mL SPE cartridges were conditioned by adding 8 to 10 mL methanol, which was
drained through the cartridge under vacuum. The methanol was then rinsed with 10 mL
deionised water twice. The samples were then loaded into the cartridges headspace free
using the Teflon tubing and the tube adapters. The flow rates were between 1 and 2
mL/min for all samples for complete passage through the sorbent. No quenching agent
was used to dechlorinate the sample, thus the cartridges were then rinsed with 20 mL
deionised water to eliminate remaining chlorine. The cartridges were then dried by
setting up a vacuum pressure of 15 mm Hg for at least 20 minutes.
The samples were then eluted with 2 mL of MtBE into a 12 mm glass test tube. No
vacuum was applied to this phase as it is important that the eluent spread throughout the
cartridge. Manual pressure using a syringe was applied after 1 hour of contact with
MtBE, to accelerate slightly the process of elution. The top layer of the test tube (MtBE)
was then transferred to a 2 mL autosampler vial capped with a PTFE septum for
analysis by GC/ECD.
Conditioning of the Strata DVB cartridge using 8 to 10 mL methanol
Cartridge rinse with 20 mL deionised water
Load the water sample into cartridge (total volume sample = 100 mL)
SPE extraction under vacuum using Teflon tubing and tube adapters, ≈1 hour, flow
rate 1-2 mL/min
Cartridge rinse with 20 mL deionised water
Back elution with 2 mL of MtBE into a test tube
Transfer final extract to conical autosampler vial
Analysis by GC/ECD
Figure 4.1 Summary of the SPE – GC/ECD method used for concentrating 16 DBPs in
treated water (adapted from Weinberg et al., 2002 and Chinn et al., 2007)
81
4.2.1.5 Instrument conditions
Two gas chromatographs, using the same instrumental conditions, were used:

A Hewlett Packard 5890 with a VF-1ms (30 m x 0.25 mm x 1 µm) column and

A Hewlett Packard 6890 with a ZB-1701 (30 m x 0.25 mm x 1 µm) column.
A volume of 1 µL was injected splitless with the injector set at 180°C. The carrier gas
was helium, set to a constant flow of 1 ml/min. An oven temperature program was used
to maximise resolution of analytes. The GC oven temperature program was as follows:
column temperature at 37°C, hold for 21 minutes, increase column temperature to
136°C at a rate of 5°C per minute and hold for 3 minutes, then finally increase column
temperature to 250°C at a rate of 20°C per minute and hold for 3 minutes. The total run
time was 52.50 minutes. The ECD was set at a temperature of 300°C. The make-up gas
was nitrogen.
4.2.2
The SPE method was used for determining the presence of 12 semi-volatile DBPs plus
four THMs in UK treated water. No calibration and quantification were attempted;
hence, no method validation was carried out.
Initially, the THMs Calibration Mix, the 551B Halogenated Volatiles Mix, TCA,
DBNM, DCIM and BCIM, all solutions at 100 µg/L in ultra pure water, were run with
the GC/ECD fitted with the VF-1ms column. The orders of DBP elution for the THMs
Calibration Mix and the 551B Halogenated Volatiles Mix (Figure 4.2) were determined
by comparison to the confirmed orders of elution of the same compounds from the US
nationwide DBP occurrence study by Weinberg et al. (2002), who used a DB-1ms
(Figure 4.3). The two columns, VF-1ms and DB-1ms, have the same polarity (1ms –
100% dimethylpolysiloxane) but are produced by different manufacturers (DB - Agilent
or VF - Varian). As seen below, the order of elution for the four THMs was TCM,
BDCM, DBCM and TBM (Figure 4.2 A, Figure 4.3) and for the 551B Halogenated
Volatiles MIX, the order was TCAN, DCAN, 1,1-DCP, TCNM, BCAN, 1,1,1-TCP and
DBAN (Figure 4.2 B, Figure 4.3).
82
Figure 4.2 Order of elution of THMs (A) and 551B Halogenated Volatiles Mix (B)
using the VF-1ms column (standard at 100 µg/L each – present study)
Figure 4.3 DB-1ms column performance using full DBP set from Weinberg et al. (2002)
(16 targeted compounds + compounds not studied here)
Running the standards altogether showed that there were two co-eluting peaks; it was
clear that BDCM and DCAN were not separated with the VF-1ms column (Figure 4.4
A), whereas they were with the DB-1ms column used by Weinberg et al. (2002) (Figure
4.3). This is believed to be due to the efficiency of the column; the column here had
already been used for a few years and the inside stationary phase may have lost its
efficiency. Hence working with the VF-1ms column, it was not possible to distinguish
between DCAN and BDCM in the standards. Therefore it was decided to run the
individual solutions on a second GC/ECD fitted with a ZB-1701 ((14%-cyanopropylphenyl)-methylpolysiloxane) column, which was available at UNC and can also be used
for the quantification of HAAs (US EPA, 2003). The detection of BDCM and DCAN at
two different retention times was possible (Figure 4.4 B). Nevertheless, the use of the
83
ZB-1701 column did not allow the detection of aldehydes; hence, it was not possible to
detect DCA and TCA, but the use of the two columns was necessary to determine the
presence of all of the compounds.
Figure 4.4 Elution of BDCM and DCAN with VF-1ms (A) and ZB-1701 (B)
Samples treated with SPE were run with both GC/ECD systems. The choice of three
days for the chlorination of the two matrices was chosen to allow enough contact time
with chlorine to form the optimum DBP concentration. We must also consider that the
literature reported degradation of the DBPs with increasing reaction time with excess
chlorine (Gurol et al., 1983; Reckhow and Singer, 1985; Ueno et al., 1996). Gurol et al.
(1983) observed that 10% of 1,1,1-TCP degrades to chloroform during a contact time of
60 minutes. Reckhow and Singer (1985) found that 1,1-DCP degrades to form 1,1,1TCP with excess chlorine, and Ueno et al. (1996) reported that the chlorination of
DCAN resulted in the formation of HAAs (DCAA and TCAA) and TCM after 60
minutes contact time with chlorine. All the targeted compounds were detected here in
the two matrices after three days of chlorination. It is noted that no quench agent was
used to stop the reaction with chlorine and this was to avoid any artifact, such as further
DBP degradation, associated with quenching agent and holding time (Gonzalez et al.,
2000; Chinn et al., 2007). Samples were extracted directly at the end of the 72 hours
incubation time.
When comparing with LLE, SPE eliminates the emulsion problems that can occur
during LLE and used less solvent (Chinn et al., 2007). However, further work here will
focus on treated water with low concentrations of organic matter and colour; hence it
84
was unlikely that a problem of emulsion would occur. The purpose of this work with
SPE was to concentrate up the DBPs to determine their presence, but LLE has been
reported to have better extraction efficiencies (Chinn et al., 2007). Moreover, LLE was
cheaper, faster and was ultimately used for further analysis. Therefore no further work
was undertaken with SPE.
4.2.3
Conclusion
The SPE method was used here as a preliminary study to determine the presence of 12
semi-volatile DBPs plus four THMs in two UK treated waters and this was successfully
achieved. However it is believed that the LLE method could be faster and cheaper than
the SPE method. Hence, further work to quantify the concentration of DBPs would use
LLE.
4.3
LLE WITH GC/ECD METHOD VALIDATION
After determining the presence of 12 selected semi-volatile DBPs plus four THMs using
SPE, the validation of LLE with analysis with GC/ECD is presented here. By
comparison with the US nationwide occurrence study by Weinberg et al. (2002), it is
expected the level of the 12 semi-volatile DBPs to be in the low range of µg/L and LLE
is believed to be a rapid and cheap method that allows detection of DPBs at low µg/L
levels.
4.3.1
4.3.1.1 Reagent and glassware
The solvent used for the LLE was MtBE, HPLC grade, the internal standard was 99%
pure 1,2-dibromopropane and sulphuric acid was used to lower the pH. Copper sulphate
was 99.995% pure, and sodium sulphate was 99% extra pure anhydrous and was baked
overnight at 100°C before the extraction. All these chemicals were provided by Fisher
Scientific (UK). The standards, their purity and their concentration were similar to these
of the Section 4.2.1.1 Table 4.1, except that some of the suppliers differ. The
85
Trihalomethanes Calibration Mix, the EPA 551B Halogenated Volatiles Mix and the
TCA were available from Sigma-Aldrich Ltd (UK). DBNM, DCIM and BCIM were as
stated in Table 4.1. DCA was available from TCI Europe (Belgium). The stock
solutions were prepared in acetonitrile (Fisher Scientific, UK) and MtBE. Ascorbic acid
(Sigma-Aldrich Ltd, UK) was used as a quench agent and has been shown to not
degrade any of the 16 DBPs (Chinn et al., 2007). Reagent water used in the preparation
of calibration standards and blanks was prepared using PURELAB Ultra Genetic 18.2
MΩ-cm pure water (ELGA LabWater, UK).
Initial stock solutions and stock solution 1 at 40 mg/L were prepared as described in
Section 4.2.1.2. Stock solution 2 at 4 mg/L was prepared by diluting exactly 1 mL of
each solution into a 10 mL volumetric flask containing MtBE. Calibration standards at
0.5, 3, 7, 15, 50 and 100 µg/L were prepared by further dilution of the stock solution 2
in 100 mL volumetric flasks with pure water. Stock 2 and calibration standards were
prepared freshly and discarded after use.
The extraction method was adapted from Krasner et al. (2001) and optimised the US
EPA Method 551.1 (1995a). It is described in Figure 4.5.
A 30 mL sample was transferred to a 60 ml glass vial, then adjusted to a pH of 3.5 or
less and extracted with 3 mL of MtBE containing an internal standard. The pH was
checked with pH paper before the addition of MtBE. The solvent phase containing the
DBPs was separated from the aqueous phase by addition of 10 g of sodium sulphate and
1 g copper sulphate. Then the sample was shaken manually for 3 to 5 minutes. The
86
layers were allowed to settle for 5 minutes and the top layer finally transferred to an
autosampler vial and analysed with GC/ECD.
Transfer 30 mL of water sample into a 60 mL glass vial with PTFE lined cap
Acidification to pH ~ 3.5 with 7/8 drops 1 M sulphuric acid solution
Add exactly 3 mL of MtBE with internal standard (IS)
IS is 100 µg/L of 1,2-dibromopropane
Add approximately 10 g muffled sodium sulphate
Add approximately 1 g copper sulphate
Shake manually for approximately 3 to 5 minutes
Transfer 1 mL of upper layer to GC vials
Analysis GC/ECD
Figure 4.5 Summary of the LLE with GC/ECD method (adapted from Krasner et al.,
2001)
Optimisation of the US EPA Method 551.1 (1995a)
The increased MtBE:water ratio in the method here (0.1:1 mL/mL) as compared to the
Method 551.1 MtBE:water ratio (0.06:1 mL/mL) is believed to offer an advantage for
the extraction of analytes with low partition coefficients (Gonzalez et al., 2000). Copper
sulphate not used in the US EPA Method 551.1 (1995a) enhanced analyte recovery and
aided in the extract transfer process (Weinberg et al., 2002).
87
4.3.1.4 Instrument conditions
DBP standards and samples were run with an Agilent 6890 GC/ECD. A volume of 1 µL
was injected splitless with the detector set at 200°C. Separation was performed by a ZB1ms column (30 m  0.25 mm  0.25 m) with a helium carrier gas at a column flow
rate of 1.0 mL/min. The initial oven temperature was 35°C and held for 22 minutes
followed by a 10°C per minute temperature ramp to 145°C and held for 2 minutes and a
final ramp of 20°C per minute ramp to 225°C and held for 10 minutes. The total run
time was 49 minutes. The detector temperature was 290°C and the rate of collection
data was 20 Hz.
4.3.1.5 Method performance
The average retention time for each DBP was determined with seven replicates of a
derived 1.0 µg/L of each DBP and internal standard aqueous standard. The retention
time precision was evaluated by calculating the relative standard deviation (RSD) as
recommended by US EPA Method 551.1 (1995a). The RSD of the sample replicate
analyses should be less than 15% and was calculated using the Equation 3.1 (Chapter
3).
Seven replicates of a derived 0.10 µg/L and 0.50 µg/L of each DBP aqueous standard
were extracted and analysed over a period of three days for determining LODs.
Ascorbic acid, which was the DBP preservative agent, was added to the standards at the
same concentration as that used to quench the samples. LOD was calculated using the
Equation 3.2 (Chapter 3). Blanks were not subtracted when performing LOD
calculations.
The analyte recovery was determined by analysing two treated water samples spiked
with three levels of DBPs. Fortifications were at 5, 25 and 75 µg/L. The recovery for
each analyte was calculated using the Equation 3.3 (Chapter 3).
4.3.1.6 Water samples
The water samples are treated lowland and upland water as described in Section 4.2.1.3.
88
4.3.2
Before carrying out the quality controls prior to analyte quantification, the retention
time of each solute was determined. As shown in the chromatograms (Figure 4.6), TCA,
DCA, DBNM, BCIM and DCIM were available individually and their retention times
were straightforward. On the other hand, THMs and the remaining species were
available as mixes. By comparison with the US EPA Method 551.1 (1995a) and the US
nationwide DBP occurrence study (Weinberg et al., 2002), which showed the order of
elution of each species encompassed in both mixes using a DB-1ms column, the
retention time identification was possible without further work.
Figure 4.6 Individual chromatograms for semi-volatile contaminants and THMs (100
µg/L each DBP)
89
A typical chromatogram for 16 DBPs and internal standard were obtained by running a
standard under the chromatographic conditions described in Section 4.3.1.4 (Figure
4.7). It has been found previously that BDCM and DCAN coeluted with a VF-1ms
column. Here the ZB-1ms column used was relatively new and all the species were
better separated. It should be noted that again BDCM and DCAN were not completely
separated but were sufficiently separated to allow quantification.
Figure 4.7 GC/ECD chromatogram for an extracted DBP16 standard (100 µg/L) (see
Table 4.2 for analyte identification)
The retention times for all 16 DBPs and internal standard are listed in Table 4.2. The
RSD calculated for the retention times reported here were low, demonstrating good
reproducibility of the method. To the knowledge of the author, this is the first published
RSD data concerning the retention times of these compounds.
90
Table 4.2 Identification of the compounds and method performance
Peak
no.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
IS
Compound
TCM
DCA
TCAN
DCAN
BDCM
TCA
1,1-DCP
TCNM
BDCM
BCAN
DCIM
1,1,1-TCP
TBM
DBAN
BCIM
DBNM
ISe
Average
tra (min)
7.72
7.95
9.99
11.17
11.24
12.02
13.12
18.53
19.10
20.01
22.64
25.28
27.10
27.71
28.28
28.81
25.72
RSDb
(%)
0.058
0.052
0.042
0.060
0.063
0.035
0.030
0.059
0.045
0.042
0.067
0.013
0.037
0.009
0.015
0.012
0.008
R2
0.999
0.993
0.992
0.997
0.995
0.994
0.998
0.995
0.997
0.994
0.996
0.995
0.999
0.944
0.996
0.998
NAf
Fortification
level (µg/L)
0.50
0.50
0.10
0.10
0.10
0.50
0.10
0.10
0.10
0.10
0.10
0.10
0.10
0.10
0.10
0.10
NAf
Detection
Limitc (µg/L)
0.088
0.124
0.020
0.019
0.036
0.029
0.029
0.039
0.049
0.023
0.086
0.089
0.095
0.014
0.108
0.059
NAf
MRLd
(µg/L)
0.264
0.371
0.061
0.057
0.108
0.086
0.086
0.117
0.148
0.070
0.257
0.268
0.284
0.041
0.324
0.178
NAf
a
The average retention time corresponds to the average of seven injections; b Corresponds to the relative
standard deviation and must be less than 15% according to US EPA Method 551.1 (1995a); c Fortified
waters were extracted and analysed over 3 days for seven replicates; d Corresponds to the minimum
reporting level and is the threshold expected for accurate quantification in an unknown sample. It has to
be at least three times the limit of detection; e 1,2-dibromopropane; f Not applicable.
To determine whether pure water could be used as a matrix for the calibration and
validation work, analyte absolute recovery was calculated in pure water, lowland and
upland water, using the calculation as in Section 4.3.1.5, Equation 3.3. Analyte absolute
recovery is based on the true response (i.e. y axis of chromatogram) and does not
involve calibration or quantitation. Thus analyte absolute recovery was assessed in pure
water and treated waters, by spiking 30 µg/L of DBPs in each matrix. Absolute
recoveries were low (40 to 60% for all matrices), but the RSD of the recovery between
matrices was less than 10%. Thus pure water was used for calibration work.
Calibration was performed by extracted procedural standards, i.e. fortified pure water,
by the procedure described in Section 4.3.1.3. Although the US EPA Method 551.1
(1995a) recommended five calibration standards, here it was decided to work with six
calibration points. Some of the compounds analysed here are not encompassed in the
US EPA Method 551.1 (1995a), and their behaviour at low range and high range of
concentrations was not known. Six calibration standards enabled coverage of low and
high concentrations and were 0.5, 3, 7, 15, 50 and 100 µg/L. The calibration curves
91
were generated by plotting area ratios (ADBP/AIS) against the concentration CDBP of the
six calibration standards (Figure 4.8)
where: ADBP = the peak area of the DBP
AIS = the peak area of the internal standard
CDBP = the concentration of the DBP.
A linear calibration curve was obtained for each of the 16 DBPs (Figure 4.8) and no
DBP was detected in pure water free of DBP spikes. The coefficients of determination
of linear regression are reported in Table 4.2 and were all above 0.944.
12
60
DCIM
TBM
BCIM
TCM
50
Area DBP/Area IS
Area DBP/Area IS
10
DCA
BDCM
TCA
1,1-DCP
TCNM
DBCM
A
8
6
4
2
40
B
30
20
10
0
0
0
20
40
60
DBP concentration (µg/L)
80
100
0
20
90
Area DBP/Area IS
70
60
60
80
100
C
TCAN
DCAN
BCAN
1,1,1-TCP
DBAN
DBNM
80
40
50
40
30
20
10
0
0
20
40
60
80
100
Figure 4.8 Calibration curves for (A) DCIM, TBM, BCIM and TCM – (B) DCA,
BDCM, TCA, 1,1-DCP, TCNM and DBCM – (C) TCAN, DCAN, BCAN, 1,1,1-TCP,
DBAN and DBNM – Concentration: 0.5, 3, 7, 15, 50 and 100 µg/L
92
The LOD (Table 4.2) for the 16 DBPs were between 0.014 and 0.124 µg/L. Detection
limits for TCM and DCIM were 0.088 and 0.086 µg/L respectively and were
comparable to a study by Gonzalez et al. (2000), where they found LOD of 0.07 and
0.09 µg/L respectively in a mixture of DBPs. The 14 remaining DBPs showed better
LOD than these reported by the same study. The US EPA Method 551.1 (1995a)
showed better recovery for every species, with LOD between 0.001 and 0.055 µg/L for
12 species (4 THMs, 4 HANs, 2 HKs, TCNM and TCA). The difference is believed to
be due to the extent of replication. Here seven replicates were used for determining
LOD, whereas eight replicates were utilised in the US EPA Method 551.1 (1995a) and
six replicates were used in the study of Gonzalez et al. (2000), modifying the Student t
value used in the calculation of the LODs. The total LOD here was 0.907 µg/L. In
relation to LOD, the minimum reporting level (MRL) for each DBP was calculated
(Table 4.2). The MRL, which is at least three times the LOD, corresponds to the
threshold expected for accurate quantification in an unknown sample. Each DBP could
be well quantified below 0.4 µg/l. It was decided to accept MRL below 0.5 µg/L
considering the good accuracy of the calibration curves.
93
Analyte recoveries exhibited bias in both water samples (Table 4.3 and Table 4.4). The
guideline provided by the US EPA Method 551.1 (1995a) recommends that recoveries
should range between 80 and 120%. However, it was decided here to accept 70 to 130%
recovery for the high levels of fortification, and 50 to 150% for the low level of
fortification, as these are the acceptable values for HAAs (US EPA, 2003). Furthermore,
the preparation of the samples here differs from the US EPA Method 551.1 (1995a), and
the method exhibited a broader range of recovery which is explained later in the text.
Table 4.3 Spike recovery in lowland water fortified with DBPs
Compounds
THMs
TCM
BDCM
DBCM
TBM
HANs
DCAN
TCAN
BCAN
DBAN
HKs
1,1-DCP
1,1,1-TCP
HNMs
TCNM
DBNM
HAs
DCA
TCA
i-THMs
DCIM
BCIM
a
Mean
RSDa (%)
recovery
(n = 3)
(%)
Mean
RSDa (%)
recovery
(n = 3)
(%)
Mean
RSDa (%)
recovery
(n = 3)
(%)
UTR c
73
85
113
UTR c
5.8
3.5
1.8
UTR c
75
81
94
UTR c
5.6
4.8
3.5
162 b
92
91
93
102.7
3.0
3.2
2.2
73
10 b
84
72
4.4
60.0
1.0
3.3
74
8b
81
78
6.0
116.0
3.0
4.0
83
17 b
88
92
12.0
83.5
5.5
3.8
86
57
1.1
3.2
83
57 b
2.1
10.4
88
60 b
1.2
31.0
37 b
54
50.2
49.1
61 b
74
18.3
13.5
73
90
22.0
8.0
ND d
ND d
ND d
ND d
ND d
ND d
ND d
ND d
48 b
66 b
5.5
11.7
124
8.7
95
10.1
86
6.5
127
10.6
90
16.0
86
9.8
551.1 (1995a); b Corresponds to biased values; c Unable to resolve; d Not detected.
94
Table 4.4 Spike recovery in upland water fortified with DBPs
Compounds
THMs
TCM
BDCM
DBCM
TBM
HANs
DCAN
TCAN
BCAN
DBAN
HKs
1,1-DCP
1,1,1-TCP
HNMs
TCNM
DBNM
HAs
DCA
TCA
i-THMs
DCIM
BCIM
Mean
RSDa (%)
recovery
(n = 3)
(%)
Mean
RSDa (%)
recovery
(n = 3)
(%)
Mean
RSDa (%)
recovery
(n = 3)
(%)
UTR c
74
83
110
UTR c
12.2
4.6
3.6
UTR c
70
81
94
UTR c
14.3
5.0
3.4
99
89
92
93
72.6
7.3
5.2
3.6
76
45 b
87
72
4.4
24.6
12.8
6.6
84
56 b
87
80
2.3
16.8
1.1
3.4
93
83
96
94
1.4
14.2
2.9
3.8
86
74
2.2
2.3
86
80
1.0
1.0
92
92
1.7
2.0
42 b
15 b
43.3
4.0
68 b
62 b
11.7
18.6
89
87
8.4
10.9
42 b
55
24.2
13.4
47 b
67 b
4.0
1.5
61 b
84
11.1
4.0
123
9.7
93
8.0
87
10.0
120
19.2
89
18.4
87
12.5
a
551.1 (1995a); b Corresponds to biased values; c Unable to resolve.
The US EPA Method 551.1 (1995a) recommends adjusting the pH of the matrix to
between 4.5 and 5.5 when collecting samples before the extraction outlined in Figure
4.5. Here the pH of both waters has not been adjusted before the extraction and was
slightly basic and neutral (7.9 and 6.7 for lowland and upland water respectively). As a
result, TCAN was biased in both waters, with poorer recovery in the lowland water.
TCAN can undergo base-catalysed hydrolysis at pH higher than 5.5 (Croué and
Reckhow, 1989), and this is believed to be the cause. 1,1,1-TCP and TCNM were also
found to be biased in the lowland water and this is mainly attributed to the matrix effect.
HNMs were biased in the upland water for concentration ≤ 25 µg/L. Despite their low
recoveries, HNMs were better recovered in the lowland water. Finally HAs were biased
in both waters. At levels ≤ 25 µg/L, DCA and TCA could not be detected in the lowland
water. Chinn et al. (2007) reported that TCA degraded almost completely at pH 8.3 after
95
being spiked at 30 µg/L for 95 minutes, and DCA showed degradation of 20%, whereas
at pH 3.5, DCA and TCA were stable for more than two weeks. Hence the bias
exhibited for DCA and TCA is likely to be caused by the initial pH of the water.
Although it is also thought that the matrix will have an impact.
It has been demonstrated by the US EPA Method 551.1 (1995a) and Gonzalez et al.
(2000) that the extraction procedure gives good recoveries when pH is below 5.5. Hence
no further work was undertaken here for analyte recovery.
TCM could not be included when the concentration was ≤ 25 µg/L due to contamination
(Figure 4.9). The contamination peak had a similar response to the TCM spike response
at 5 and 25 µg/L, making the recovery percentage impossible to determine. Furthermore
recovery observed at 75 µg/L exhibited bias and/or very high RSD and this is believed
to be due to the contamination and not the matrix impact. Indeed previous work
(Parsons et al., 2009) assessing the recovery of TCM in the same waters with the US
EPA Method 551.1 (1995a) found values of 86 to 134% at low levels, concluding no
bias for TCM in both waters. The origin of the contamination has not been determined.
Nevertheless, because LOD could be detected and a calibration curve could be drawn,
no further work was undertaken for analyte recovery of TCM.
Figure 4.9 Contamination and TCM peaks while working on recovery in lowland water
96
4.3.3
Limitations
Despite rapid sample preparation and good limits of detection, a few parameters must be
considered while determining the level of the 12 semi-volatile DBPs plus four THMs
with LLE. First of all it is recommended to acidify the samples as soon as the reaction is
finished and to extract and process the samples as soon as they have been quenched to
avoid any possible degradation due to base-catalysed hydrolysis. The quenching agent
can also be responsible for degradation (Gonzalez et al., 2000) and here we quenched
with ascorbic acid, which has been shown not to degrade the 16 DBPs (Chinn et al.,
2007). However, it is recommended for further work that could include more DBPs,
such as bromopicrin or tribromoacetonitrile, etc. to investigate the use of ammonium
chloride as secondary quenching agent because these DBPs are not stable in the
presence of ascorbic acid (Chinn et al., 2007).
4.3.4
Conclusion
The method validated here was found to be both reproducible and precise with some
bias. Linear calibration curves were reported for each DBP. Spike recovery calculations
showed biased analytes in both waters and this was attributed to the matrix impact and
the initial pH of the water samples. Hence the investigation of accuracy, precision,
working range, selectivity showed the method to be reliable and the limit of detection of
0.17 µg/L for each DBP was met, validating the method for further work.
4.4
CHAPTER SUMMARY
The results presented here showed that 12 semi-volatile DBPs, plus four THMs were
present in two different UK treated waters, when concentrated by a factor of 50 using
SPE. Nevertheless, the LLE method requires less time for the sample preparation and it
is cheaper compared to SPE. Hence, LLE was chosen for the quantification of these 16
DBPs. Low µg/L concentrations of these compounds in treated water were expected and
the LLE method allowed quantification at these levels. In addition LOD, detected for
each DBP, was found to be satisfactory, validating the method. Bearing in mind that
limitations of the method concern the pH and the quenching agent, it is recommended to
97
lower the pH to 4.5 – 5.5 to avoid any possible DBP degradation. It is also important to
choose a quenching agent that is inert toward the DBPs.
98
Chapter 5
COMPARISON OF THE DBP-FP OF UK TREATED
WATERS EXPOSED TO CHLORINE AND
MONOCHLORAMINE
5.1
INTRODUCTION
Drinking water DBPs result from the reaction of chlorine NOM and/or bromide/iodide
present in drinking water supplies (Rook et al., 1974). THMs are the only regulated
DBPs in the UK and it is required by law that the sum of four THMs does not exceed
100 µg/L with a frequency of sampling dependent on the population size. HAAs are
often found to be as prevalent as THMs but are currently not regulated in the UK.
However, the European Union is considering regulating the nine HAAs at 80 µg/L
(Cortvriend, 2008) and as such there is growing interest in the levels of these
compounds found in UK drinking waters and how best to control them.
In order to comply with these proposed regulations, there has been an increasing interest
in using monochloramine as a secondary disinfectant because of reduced DBP
formation and its ability to provide residuals in water distribution systems.
Monochloramine is known to only form trace amounts of THMs and HAAs. However,
the formation of DXAAs, although, generally lower than with chlorine, can still reach
significant levels depending on the dose, chlorine to ammonia ratio, pH and other
conditions (Diehl et al., 2000; Hua and Reckhow, 2007). The use of monochloramine
may also lead to an increase in other DBPs such as HANs and i-THMs (Krasner et al.,
1989; Bichsel and Von Gunten, 2000). HANs and i-THMs are two unregulated classes
of semi-volatile DBPs also present in disinfected waters alongside other unregulated
DBPs including HNMs, HAs and HKs (Krasner et al., 2006). These semi-volatile DBPs
are of interest because of their toxicity. HANs have been reported to be genotoxic and
potentially carcinogenic for human health and HKs exerted carcinogenic or mutagenic
effects in mice (Bull and Robinson, 1986; Daniel et al., 1986). Plewa et al. (2004) found
HNMs to be toxic in CHO cells and Richardson (2003) suggested that i-THMs
99
Chapter 5 – Comparison of DBP-FP in UK drinking water exposed to Cl2 and NH2Cl
could be more toxic than their brominated and chlorinated analogues.
Past research has established that levels of HAAs and THMs in chlorinated waters vary
according to the levels of their precursors. High NOM concentrations have generally
been associated with high HAA and THM concentrations (Liang and Singer, 2003;
Sharp et al., 2006a) and nitrogenous precursors from algae or effluent organic matter
(EfOM) have been related to nitrogenous DBPs, such as HANs (Oliver et al., 1983).
The presence of bromide in water will also affect the concentration of DBPs as will
other factors such as the disinfectant dose applied, the pH, the temperature of the water
samples and the reaction time of disinfectant in water (Singer et al., 2002). To better
control and understand the formation of DBPs in water samples, the use of FP tests have
been widely used (Zhang et al., 2000; Liang and Singer, 2003; Ates et al., 2007;
Krasner et al., 2007). FP tests are usually conducted with controlled pH, controlled
temperature and relatively high chlorine concentration dosed for a long contact time in
order to maximise DBPs formation (Krasner et al., 2007).
Because HAAs are not routinely measured in the UK, there is currently little
information concerning their levels, their formation route and their relation to other
DBPs. Furthermore an insight into semi-volatile DBPs was of interest, because of their
toxicity and their presence in disinfected waters using chlorine or monochloramine
(Krasner et al., 2006). To have a better understanding of HAAs, THMs and semivolatile DBPs in UK treated waters, their formation was evaluated under controlled
conditions. The use of FP tests allowed a direct comparison between the use of chlorine
and monochloramine and the impact the change of disinfectant had on the DBP levels.
The objectives of the work presented in this chapter were (1) to determine the relative
distribution and speciation of DBPs from treated water using FP tests, (2) to compare
the formation of DBPs with chlorine or monochloramine and (3) to identify any
relationships between water sources and DBPs. Here 11 WTWs across England and
Wales have been surveyed to give a wide range of water sources as well as different
water characteristics. The DBPs included THM4, HAA9 but also four HANs (DCAN,
TCAN, BCAN and DBAN), two HKs (1,1-DCP and 1,1,1-TCP), two HAs (DCA and
TCA), two HNMs (TCNM and DBNM) and two i-THMs (DCIM and BCIM).
100
5.2
MATERIALS AND METHODS
5.2.1
Water sample collection
Treated water samples were collected in July 2008 from 11 WTWs (Table 5.1), selected
by the water companies and spread geographically across England and Wales. Samples
were collected at the end of treatment processes but prior to disinfection. The processes
from each works are outlined in Table 5.1. All samples were collected in polyethylene
or glass 1L bottles and shipped to Cranfield laboratories. All samples were kept
refrigerated at 5°C until analysis. Analyses were conducted within 7 to 15 days after the
samples were received.
5.2.2
Water sample characterisation
5.2.2.1 pH
pH was measured with a Jenway 3310 pH meter, calibrated each day of use with buffer
solution pH 4 (phthalate), pH 7 (phosphate) and pH 10 (borate) (Fisher Scientific, UK).
5.2.2.2 Non purgeable organic carbon (NPOC)
NPOC was measured using a Shimadzu TOC-5000A analyser (Shimadzu, Milton
Keynes, UK). Samples were acidified and purged with air to convert the inorganic
carbon to CO2. The total carbon (TC) was then measured and was referred to as NPOC.
The TC standard was made by dissolving 2.125 g potassium hydrogen phthalate in 1L
ultra pure water.
The standard had a concentration of 1000 mg/L and working standards were diluted to
the appropriate concentration with pure water. The machine was calibrated on the day of
analysis. The analyser took up to five replicates and reported an average of three given
that the coefficient of variance was not greater than 2%. The analyser was recalibrated if
the value of the standards were not within 10% of the expected value.
101
5.2.2.3 UV / SUVA
UV absorbance at 254 nm was measured using a Jenway 6505 UV/VIS
spectrophotometer (Patterson Scientific Ltd., Luton, UK), calibrated using pure water as
a blank.
SUVA (L/m-mg C) was calculated as the ratio of UV absorbance at 254 nm (1/m) to
NPOC (mg C/L).
5.2.2.4 Bromine and iodine measurement
Samples were analysed using an Elan 9000 inductively coupled plasma mass
spectrometry (ICP/MS) from Perkin Elmer (UK). The ICP/MS detects only elemental
ions. The concentration of each element is determined by comparing the counts for a
selected isotope to an external calibration curve generated for that element. For
bromine, a calibration curve was set up from 10 to 500 µg/L and for iodine, 0.5 to 20
µg/L. Liquid samples were introduced to the ICP/MS by a peristaltic pump whereupon
samples were nebulised and sprayed into the instrument to meet the high temperature
plasma. For each sample, three replicates were analysed at a rate of 60 sweeps per
reading. The integration time was set at 3000 ms and the dwell time 50 ms per atomic
mass unit. The scan mode was peak hopping and the carrier gas was argon set at 1
mL/min.
5.2.3
DBP-FP
5.2.3.1 Chlorine and monochloramine solution
Preparation and determination of the chlorine solution was carried out using the
procedure 4500-Cl B. Iodometric Method I in “Standards Methods for the Examination
of Water and Wastewater” (APHA, 1992).
Preparation and determination of the monochloramine solution was carried out using the
method 4500-Cl F. DPD Ferrous Titrimetric Method (APHA, 1992).
102
5.2.3.2 Buffer
A stock solution of sodium phosphate dibasic (Na2HPO4, Fisher Scientific, UK) at 1/15
M and potassium acid phosphate (KH2PO4, Fisher Scientific, UK) at 1/15 M was
prepared respectively by dissolving 4.733 g in 0.5 L of ultra pure water and 4.540 g in
0.5 L of ultra pure water. A buffer at pH 7.2 was making up by adding 27 mL of the
KH2PO4 stock solution to 73 mL of Na2HPO4. The buffer was made fresh when the pH
fell below 0.2 pH units of the expected value.
5.2.3.3 Quench solutions
Ammonium chloride at a concentration of 100 mg/L was used to quench chlorine and
monochloramine residual while not degrading HAAs, in particular HAA3 (BDCAA,
DBCAA and TBAA) (Singer et al., 2002). This solution was discarded after two weeks.
Ascorbic acid at a concentration of 35 mg/L was used to quench chlorine and
monochloramine residual in THM and semi-volatile DBP samples. The choice is based
on the fact that ascorbic acid has been shown not to degrade any of these 16 DBPs
(Chinn et al., 2007). This solution was discarded after two weeks.
HAA, THM and the semi-volatile DBP samples were extracted immediately after
incubation time to avoid possible artifacts associated with quenching agents and holding
times.
5.2.3.4 Samples exposed to chlorine and monochloramine
For DBP-FP tests conducted in the presence of chlorine, samples were chlorinated at pH
7.2; based on the NPOC level in each sample, chlorine:NPOC ratio was 3:1 on a weight
basis. A 100 mL bottle was partly filled with the water sample, the buffer and the
chlorine solution were added and the bottle was filled completely and capped headspace
free with a PTFE-lined cap. Samples were placed for 24 hours at 20°C in the dark. At
the end of the incubation period the chlorine residual was quenched with 136 µL of
ammonium chloride for HAA samples and with 50 µL of ascorbic acid for the THM and
the semi-volatile DBP samples.
103
DBP-FP tests conducted in the presence of monochloramine were carried out using
preformed monochloramine created by mixing aqueous ammonium sulphate and
sodium hypochlorite solutions (HOCl) in accordance with Diehl et al. (2000). A
chlorine to nitrogen mass ratio of 3:1 was used in all samples and addition of
monochloramine was based on the NPOC level, with combined chlorine:NPOC ratio of
3:1 by weight. The procedure of monochloraminated samples was the same as that for
chlorinated samples and the quench agents were identical.
5.2.4
Sample analytical methods
HAA samples were extracted with a method reported by Tung et al. (2006) and is a
modified version of US EPA Method 552.2 (1995). The method description and the
instrumentation conditions are reported in Chapter 3 Section 3.3.
THMs and the semi-volatile DBPs were extracted with an adapted method from Krasner
et al. (2001). The method description and the instrumentation conditions are reported in
Chapter 4 Section 4.3.
5.3
RESULTS AND DISCUSSION
5.3.1
Water characterisation
Samples of treated waters collected from drinking WTWs across England and Wales
were analysed for pH, NPOC, UV, bromine and iodine. These results are presented
along with calculated SUVA values (Table 5.1). The average NPOC concentration was
1.6 mg/L with the highest value (3.7 mg/L) found in LR and the lowest concentration
(0.2 mg/L) in B1. The NPOC concentration of the lowland rivers (mean of 1.7 mg/L)
was similar to that measured in the upland reservoirs (mean of 1.5 mg/L).
SUVA is a useful parameter when assessing NOM as it indicates the UV-absorbing
chromophores (aromatic carbon moieties) from molecules and is strongly linked to the
precursor polarity and hence DBP concentrations (Edzwald and Tobiason, 1999; Hwang
et al., 2001). SUVA values calculated here ranged from 1.5 L/m-mg C (B1) to 5.4 L/mmg C (UR3). L1 and UR3, with low NPOC values (1.2 and 1.1 mg/L respectively), had
high SUVA values of 4.6 and 5.4 L/m-mg C respectively, which indicate that the NOM
104
was hydrophobic in character. Previously, hydrophobic NOM has been reported to
preferentially form THMs and TXAAs (Hwang et al., 2001). In general, the upland
waters had greater SUVA values than the lowland waters and boreholes (average of 3.5,
1.9 and 1.7 L/m-mg C respectively).
Table 5.1 List of WTWs, sources and water characteristics
Work
Work description
ref.
B1
B2
NPOC UV254 SUVA254 Bromine Iodine
(mg/L) (1/m) (L/m-mg C) (µg/L) (µg/L)
BOREHOLE (B)
Sampling point: Post filter
Main process: Filtration
Sampling point: Post membrane
prior to superchlorination
Main process: Membrane
filtration with pre-oxidation
pH
7.8
0.2
0.4
1.5
275
3.5
7.2
1.2
2.2
1.8
42
6.9
4.6
31
1.3
2.7
75
16.7
1.6
209
8.9
2.6
44
0.9
2.4
18
0.9
5.4
29
0.9
2.4
14
0.9
1.8
310
6.3
1.7
108
3.0
LAKE (L)
L1
L2
LR
UR1
UR2
UR3
BR1
BR2
BR3
Sampling point: Post membrane
Main process: Membrane
5.9
1.2
5.5
filtration
Main process: Coagulation/Direct 6.8
1.2
3.4
filtration
LOWLAND RESERVOIR (LR)
Sampling point: Post GAC
Main process: Ozone/coagulation 7.8
3.7
5.8
/GAC
UPLAND RESERVOIR (UR)
Sampling point: Post sand
filtration
7.4
1.6
4.2
Main process: Coagulation
8.9
1.7
4.1
Main process:
Coagulation/filtration
Sampling point: Post slow sand
filter
6.2
1.1
5.9
Main process: Direct filtration
LOWLAND RIVER (BR)
5.5
2.2
5.3
Main process: Coagulation/GAC
7.5
1.6
2.9
7.2
1.4
2.4
ref. - reference
105
No specific trends were observed between the water treatment processes used and the
treated water SUVA values. The two waters with the highest SUVA (L1 and UR3) were
treated with direct filtration, not coagulation, which is more effective towards removal
of hydrophobic material (Sharp et al., 2006a).
The level of bromine, which is assumed here to be mainly bromide, ranged from 14 to
310 µg/L (Table 5.1), with an average concentration of 105 µg/L. This is in line with
the levels reported by Amy et al. (1994) who reported that typical concentrations of
bromide in natural waters ranging from 30 to 200 µg/L, with an average of 100 µg/L.
The highest concentrations were found in B1, LR, BR2 and BR3. It is expected here
that the waters with levels of bromide > 100 µg/L would form brominated DBPs (Singer
et al., 2002).
The level of iodine found during this survey varied between 0.9 and 16.7 µg/L (Table
5.1) and is in line with the findings of Fuge et al. (1986) who found total iodine in water
sources between 0.5 and 20 µg/L. Interestingly the ratio of bromine to iodine here
varied considerably between 1 and 22%, which indicates no specific trend between the
level of bromine and iodine in the water sources.
5.3.2
DBP levels from different water sources
5.3.2.1 HAAs
The concentrations of nine HAAs from the 11 treated waters were quantified after
exposure to chlorine and monochloramine (Figure 5.1). In Figure 5.1, chlorine data are
represented as the treatment work reference only (e.g. B1) and the monochloramine data
are shown as NH2Cl-work reference (e.g. NH2Cl-B1). It is clear that using
monochloramine produced significantly less HAAs (average reduction of 77%) when
compared to chlorine. These findings compare well with previous studies that have
looked at HAA formation when using preformed monochloramine (90 to 95%
reduction) (Cowman and Singer, 1996; Guay et al., 2005).
When chlorine is used as the disinfectant (Figure 5.1), considerable variation was
observed between the individual waters with HAA levels ranging from 5.0 to 69 µg/L,
with an average value of 37 µg/L. It is the first FP data available for HAAs in England
106
and Wales, although Malliarou et al. (2005), who reported their data on a regional basis,
found means of 35, 52 and 95 µg/L in finished waters from three regions investigated in
England and Wales. Here, no relationship could be identified between water source type
and HAA levels.
Across the chlorinated water samples, the major species formed were TCAA (ranging
from 1.0 to 40 µg/L) and DCAA (ranging from 2.5 to 22 µg/L). Sérodes et al. (2003)
also found TCAA and DCAA to be the major species formed in treated waters from
Quebec exposed to FP tests using chlorine. On a mass basis, DCAA and TCAA were
followed here by BDCAA, BCAA, MCAA and DBCAA. For all waters, MBAA,
DBAA and TBAA were the least concentrated, with TBAA not always detected.
The ratio of TCAA:DCAA varied across the chlorinated samples, with TCAA being
predominant in six of the treated waters (B2, L2, UR1, UR2, UR3 and BR1), and
DCAA for the remaining waters (B1, L1, LR, BR2 and BR3). This variation in ratio
was also observed by Sérodes et al. (2003). The excess chlorine used during FP tests as
well as the bromine concentration are believed to be the cause. When the bromine
concentration was ≤ 75 µg/L, TCAA was predominantly formed here. When there was a
high concentration of bromine (> 100 µg/L) (water samples B1, LR, BR2 and BR3) and
an excess of chlorine, it is believed that bromide reacted to form hypobromous acid
(HOBr/OBr-), which is known to react with NOM faster than aqueous chlorine
(Westerhoff et al., 2004). Consequently, the NaOCl to the NPOC ratio (NaOCl:NPOC),
on a mass basis, decreased as the bromine increased. Miller and Uden (1983) amongst
others found that at lower NaOCl:NPOC, the relative amount of DCAA formed was
higher than that of TCAA, which was observed here. For example BR1, with a bromine
concentration of 14 µg/L, formed 22 µg/L of DCAA and 40 µg/L of TCAA, whereas
LR, with a bromine concentration of 209 µg/L formed 16 µg/L of DCAA and 12 µg/L
of TCAA.
In addition, it was observed here that treated waters (LR, BR2 and BR3) with high
levels of bromine (> 100 µg/L) formed between 43 and 52% of brominated species of
the total HAA. Peters et al. (1991) found that brominated species accounted for 60% of
the nine HAAs in a study of highly brominated Dutch waters (bromide concentration
between 100 and 500 µg/L). Furthermore, it was observed here that in low bromine-
107
containing waters (≤ 75 µg/L), chlorinated HAAs dominated over brominated HAAs, as
previously found by Golfinopoulos and Nikolaou (2005) and Ates et al. (2007).
Neither SUVA nor NPOC were effective surrogates for predicting the formation of
HAAs in these waters. Correlations between water quality parameters and DBP-FP will
be investigated in detail later in this chapter.
MCAA
70
60
MBAA
DCAA
BCAA
TCAA
DBAA
BDCAA
DBCAA
TBAA
work ref. = chlorine (e.g. B1)
NH2Cl-work ref.= monochloramine
(e.g. NH2Cl-B1)
Concentration HAAs (µg/L)
50
40
30
20
10
BR3
Cl-B
R3
NH2
BR2
Cl-B
R2
NH2
BR1
Cl-B
R1
NH2
UR3
NH2
Cl-U
R3
UR2
NH2
Cl-U
R2
UR1
NH2
Cl-U
R1
LR
NH2
Cl-LR
L2
Cl-L2
NH2
L1
Cl-L1
NH2
B2
Cl-B
2
NH2
Cl-B
1
NH2
B1
0
Figure 5.1 Distribution of HAAs after 24 hours bench scale exposure to chlorine and
monochloramine for 11 treated waters
When monochloramine was used as the disinfectant the highest concentration of HAAs
formed was 14 µg/L (L2, LR and BR1) and the average concentration was 8.2 µg/L
(Figure 5.1). DXAAs, and in particular DCAA, were observed to be the predominant
HAAs formed in all the monochloramine water samples, comprising at least 60% of the
total HAA formation. Karanfil et al. (2008) and Cowman and Singer (1996) both
reported DXAA to be the main HAA species when using monochloramine and, in their
studies, constituted 80 and 65% respectively of the total HAA formed. The
108
concentration of MXAA was always the lowest of the HAAs measured, but did
contribute a maximum of 20% in some cases.
The difference in HAA concentrations obtained with chlorine and monochloramine is
mainly believed to be due to different formation routes. When using chlorine, it was
concluded that its reaction with NOM preferentially forms TCAA in low brominecontaining waters. However, the formation mechanistic with monochloramine is more
complex and different models have been proposed in the literature. Karanfil et al.
(2007) and Hong et al. (2007) both showed that the direct reaction between preformed
monochloramine and NOM is responsible for about 80% of HAA formation and that the
remaining HAA formation was attributed to the dissociation of monochloramine to
chlorine. Duirk and Valentine (2006) attributed the formation of DXAA to be mostly
from the reaction between NOM and chlorine in equilibrium with monochloramine. The
presence of bromide in the samples complicates the chemistry of the system because
bromide reacts with free chlorine and/or monochloramine to form HOBr/OBr-,
bromamines and bromochloramine (Valentine, 1986; Diehl et al., 2000). Here, the
concentration of TXAA, and especially TCAA remains high in many of the
monochloraminated samples, such as B1, L2, LR, BR2 and BR3, whilst in others, such
as UR1 or UR2, the main species was DCAA, making it unclear as to which mechanism
is predominant.
Bromine incorporation
It has been found that bromine played an important role in the formation of HAAs.
Indeed it was previously concluded here that:

Levels of bromine > 100 µg/L are responsible for the predominance of DCAA in
chlorinated waters.

High bromine-containing waters (> 100 µg/L) formed more brominated HAA
species than low bromine-containing waters.

Finally, bromine in waters exposed to monochloramine preferentially form
brominated DXAAs.
109
To assess the extent of bromine substitution in HAA when using chlorine and
monochloramine, the bromine incorporation factor (BIF) was calculated (Symons et al.,
1993):
BIF 
HAABr9 (µmol/L)
,
HAA9 (µmol/L)
Equation 5.1
where HAABr9 is the sum of the molar concentrations of bromine incorporated in the
nine HAA species and HAA9 represents the sum of molar concentrations of all nine
HAAs. The value BIF can range from zero to three.
Calculated BIF values were plotted against the bromine concentration (Figure 5.2).
0.6
Chlorine
Monochloramine
0.5
Chlorine
R2 = 0.39
BIF
0.4
0.3
Monochloramine
R2 = 0.72
0.2
0.1
0
0
50
100
150
200
250
300
350
Bromine (µg/L)
Figure 5.2 BIF in chlorinated and monochloraminated samples versus bromine
concentration
The results showed limited correlations between BIF and the bromine concentration.
BIF was higher in most of the chlorinated samples than in the monochloraminated
samples. Chlorine is a more powerful oxidant and its reaction with bromine to form
HOBr and then the formation of brominated HAAs will be faster and more predominant
than with monochloramine (Deborde and Von Gunten, 2008). The correlation between
BIF and bromine was better in water exposed to monochloramine (R2 = 0.72) than to
110
chlorine (R2 = 0.39). As a conclusion, the study of BIF was shown to not give
conclusive evidence to justify that increasing bromide concentrations increased BIF.
5.3.2.2 THMs and i-THMs
THMs
The THMs formed following exposure to both chlorine and monochloramine were
quantified and are presented in Figure 5.3. As with the HAAs, shifting from chlorine to
monochloramine produced significantly less THMs (average reduction of 92%). It is
noted that when monochloraminating, TCM could not be detected as its level was below
the level of a contamination peak.
While using chlorine there was as expected considerable variation in THM levels across
the 11 waters with concentrations ranging from 2.6 to 66 µg/L observed. The average
concentration was 30 µg/L, which is similar to the value observed for the HAAs
(average of 37 µg/L). The lowest concentration of THMs was found in L1 and the
highest in LR, followed by L2. These results are similar to those for the HAAs, and
specifically, the concentration of TCM was similar to that of TCAA in many samples,
indicating possible common precursors. For example, in B2, TCM was 13 µg/L and
TCAA was 11 µg/L, in UR1, both TCM and TCAA were concentrated at 25 µg/L and
in BR1, TCM was 35 µg/L and TCAA was 40 µg/L. It was observed that L1 had a
lower concentration of THMs than L2 (2.6 and 47 µg/L respectively), both waters
having the same NPOC values, but L1 having a greater SUVA value than L2, which
indicates that neither NPOC, nor SUVA were effective surrogate for these two treated
waters.
In all the chlorinated waters with bromine < 50 µg/L, TCM was found to be the major
THM species, whereas in those waters with bromine ≥ 75 g/L brominated THMs
became the major group.
When using monochloramine the concentrations of THMs were mostly below 1 µg/L,
aside from B1, LR and BR2 (Figure 5.3). Interestingly, BR2, which had the highest
concentration of bromine (310 µg/L) could form brominated THMs up to 13 µg/L while
using monochloramine as a disinfectant. When bromine is present in the water, it reacts
111
with monochloramine to form bromamine species, which are highly reactive (Diehl et
al., 2000). However the reactivity of bromamines with NOM to form THMs remains
unknown here.
TCM
70
Concentration THMs (µg/L)
60
BDCM
DBCM
TBM
(e.g. NH2Cl-B1)
50
40
30
20
10
BR3
Cl-B
R3
NH2
BR2
Cl-B
R2
NH2
BR1
Cl-B
R1
NH2
UR3
NH2
Cl-U
R3
UR2
NH2
Cl-U
R2
UR1
NH2
Cl-U
R1
LR
NH2
Cl-LR
L2
Cl-L2
NH2
L1
Cl-L1
NH2
B2
Cl-B
2
NH2
Cl-B
1
NH2
B1
0
Figure 5.3 Distribution of THMs after 24 hours bench scale exposure to chlorine and
i-THMs
The concentrations of the two i-THMs formed with both chlorine and monochloramine
are shown in Figure 5.4. The maximum concentration found here was 0.73 µg/L and
most concentrations were below the MRL of 0.58 µg/L (see Chapter 4). Cancho et al.
(2000) reported average levels lower than 1 µg/L for three species (DCIM, BCIM and
DBIM) in sand filters and ozonated waters, and Krasner et al., (2006) reported a
maximum of 19 µg/L for six i-THMs with DCIM and BCIM being the prevalent
species. Here, it was found that the BCIM was generally at higher levels than DCIM.
Overall the concentration of i-THMs formed was low when compared to THMs (Figure
112
5.4), with the ratio of the i-THMs to THMs being 1% on an average basis and 0.4% on a
median basis. Krasner et al. (2006) reported a median ratio of 2% but for six i-THMs. It
is known that chlorine oxidises iodide through to iodate (IO3-) and, hence, minimises
any potential for i-THM formation (Bichsel and Von Gunten, 1999). Thus, chlorine,
which is largely in excess here, generated mainly IO3-, which is believed to be the main
reason for the low level of i-THMs and the fact that there is no correlation between the
i-THMs and the iodine level in the water sources.
DCIM
BCIM
NH2Cl-work ref.= monochloramine (e.g. NH 2Cl-B1)
0.8
0.6
0.4
0.2
BR3
Cl-B
R3
NH2
BR2
Cl-B
R2
NH2
BR1
Cl-B
R1
ND ND
NH2
UR2
NH2
Cl-U
R2
UR1
NH2
Cl-U
R1
LR
NH2
Cl-LR
L2
Cl-L2
NH2
Cl-L1
L1
B2
Cl-B
2
NH2
Cl-B
1
B1
NH2
UR3
NH2
Cl-U
R3
ND
ND
0
NH2
Concentration i-THMs (µg/L)
1
Figure 5.4 Distribution of i-THMs after 24 hours bench exposure to chlorine and
monochloramine for 11 treated waters (ND – not detected)
Previous research has shown that the formation of i-THMs is favoured by
monochloramine (Bichsel and von Gunten, 2000), because monochloramine, unlike
chlorine, is unable to oxidise hypoiodus acid (HOI) to IO3- meaning that HOI has a
longer lifetime with monochloramine and can react with NOM to form i-THMs (Bichsel
and Von Gunten, 2000a). Here, it was found that levels of i-THMs after
monochloramine were between not detected to 0.89 µg/L (Figure 5.4), with five water
samples (B1, B2, L1, L2 and BR2) having greater concentrations of i-THMs than after
exposure to chlorine, whereas the contrary was observed in LR, UR1, UR2, UR3 and
BR3. The monochloramine results presented in Figure 5.4 do not show any trend, which
is most probably due to measuring close to the limit of detection. It can also be noted
113
that one species was detected in each sample except for BR2, where DCIM and BCIM
were detected, a fact for which no explanation has been found.
As a conclusion, the use of FP tests was shown not to give conclusive evidence for the
true levels of i-THMs.
5.3.2.3 HANs
The concentrations of four HANs measured in chlorinated and monochloraminated
waters are presented in Figure 5.5. When using chlorine, HANs were detected in all
treated waters and their concentrations were usually an order of magnitude lower than
the concentrations of THMs and HAAs. Total HAN concentrations ranged between
0.023 and 5.5 µg/L, which is similar to the results of Krasner et al. (2007), who reported
levels of dihalogenated HANs between approximately 0.80 µg/L and 6.2 µg/L when
using FP tests for 24 hours. DCAN was the major HAN formed and contributed up to
56% of the total HAN, followed by BCAN (27%), DBAN (16%) and TCAN (2%).
Dihalogenated HANs are reported to be more stable than the trihalogenated HANs by a
number of studies (Coleman et al., 1984; Reckhow and Singer, 1984; Peters et al., 1990;
Singer et al., 1995; Glezer et al., 1999). In addition, TCAN can undergo base-catalysed
hydrolysis at pH higher 5.5 (here, the pH was 7.2) (Croué and Reckhow, 1989), which
explains why it was rarely detected in the samples here.
DCAN was the most abundant species measured in chlorinated waters containing levels
of bromine < 50 µg/L. In the waters with bromine  75 µg/L the brominated HANs
(BCAN and DBAN) were dominant (67% total HAN). Peters et al. (1990) reported a
similar value with the brominated dihalogenated HANs accounting for 60% of the total
HAN in Dutch surface waters, which contained bromide concentrations ≥ 500 µg/L. It
was also noticed here, that the lowland waters L1, BR2 and BR3 produced more HANs,
which could be explained by the difference in precursors between lowland and upland
waters remaining after treatment processes. Lowland sources are more likely to contain
DON, the main precursor for HANs (Oliver, 1983; Lee et al., 2007).
114
Changing from chlorine to monochloramine decreased the concentration of HANs by
81% (Figure 5.5). Hua and Reckhow (2007) also found that concentrations of HAN
were reduced by between 93% and 100% when using monochloramine and little
dihalogenated HANs (<1 µg/L) were formed.
The speciation observed was again dependent on the presence of bromine. For example
BR2, which contains 310 µg/L of bromine, formed mainly BCAN and DBAN (0.31 and
0.39 µg/L respectively), whereas UR2 with a bromine concentration of 18 µg/L formed
0.013 and 0.014 µg/L for both BCAN and DBAN, but 0.26 µg/L of DCAN. Yang et al.
(2007), who studied the precursors of HAN, observed that the nitrogen precursors of
DCAN are from monochloramine and nitrogen-containing compounds in the sample.
The study worked with isotoped monochloramine (15NH2Cl) and found that the
percentage of DCA15N in the total DCAN yields varied from 8% to 78% in
monochloraminated model compound solutions.
DCAN
6
TCAN
BCAN
DBAN
(e.g. NH2Cl-B1)
Concentration HANs (µg/L)
5
4
3
2
1
BR3
Cl-B
R3
NH2
BR2
Cl-B
R2
NH2
BR1
Cl-B
R1
NH2
UR3
NH2
Cl-U
R3
UR2
NH2
Cl-U
R2
UR1
NH2
Cl-U
R1
LR
NH2
Cl-LR
L2
Cl-L2
NH2
L1
Cl-L
1
NH2
B2
Cl-B
2
NH2
Cl-B
1
NH2
B1
0
Figure 5.5 Distribution of HANs after 24 hours bench scale exposure to chlorine and
115
5.3.2.4 HKs, HAs and HNMs
HKs
The concentrations of the two HKs formed following exposure to chlorine and
monochloramine are presented in Figure 5.6. HKs were detected in all the treated waters
exposed to chlorine (Figure 5.6), with concentrations ranging from 0.37 to 3.9 µg/L,
with a mean value of 1.8 µg/L. The highest concentration was observed in BR1,
whereas the lowest concentration was measured in L1 and B1.
1,1,1-TCP was the major HK formed in B2, L2, LR, UR1, UR2, UR3, BR1, BR2 and
BR3. The greater formation of 1,1,1-TCP in the samples is believed to be the result of
the excess chlorine used in FP test, involving the oxidation of 1,1-DCP to 1,1,1-TCP
(Gurol et al., 1983).
The use of monochloramine resulted in an average decrease of 70% in the total HK
compared to the use of chlorine (Figure 5.6); no 1,1,1-TCP was detected, which agrees
with Yang et al. (2007). Monochloramine does not provide enough free chlorine to lead
further substitution into 1,1-DCP, and Yang et al. (2007) found 1,1-DCP to be stable
with monochloramine, which explains its abundance here (average of 32% more than
with chlorine).
1,1-DCP
Concentration HKs (µg/L)
4
1,1,1-TCP
NH2Cl-work ref.= monochloramine (e.g. NH 2Cl-B1)
3
2
1
BR3
Cl-B
R3
NH2
BR2
Cl-B
R2
NH2
BR1
Cl-B
R1
NH2
UR3
NH2
Cl-U
R3
UR2
NH2
Cl-U
R2
UR1
NH2
Cl-U
R1
LR
NH2
Cl-LR
L2
Cl-L2
NH2
L1
Cl-L
1
NH2
B2
Cl-B
2
NH2
Cl-B
1
NH2
B1
0
Figure 5.6 Distribution of HKs after 24 hours bench scale exposure to chlorine and
116
HAs
The concentration of HAs following the use of chlorine or monochloramine is shown in
Figure 5.7. HAs were present in all samples after 24 hours contact time with chlorine.
The minimum value was 0.92 µg/L for L1 and the maximum value was 9.5 µg/L for
BR1. The average of HAs formed was 4.4 µg/L, hence this group of DBPs represented
the third major class of halogenated DBPs formed (on a weight basis) after HAAs and
THMs, which is in agreement with Krasner et al. (2001). The major HA detected was
TCA (also called chloral hydrate). Williams et al. (1997) also found TCA to be the most
prevalent DBP after HAAs and THMs and Koudjonou et al. (2008) reported TCA in
drinking water made up 60% of the total HA. It has been shown previously that
ozonation can increase the levels of DCA and TCA (Weinberg et al., 1993). Here, the
two boreholes B1 and B2 had different concentrations of HKs, with B2, the preozonated site, having a greater formation potential for DCA (0.62 µg/L) and TCA (2.4
µg/L) than B1 (0.31 and 0.61 µg/L respectively), which has no ozone.
The use of monochloramine resulted on average in a 90% decrease in the total HA
concentration (Figure 5.7) and as observed by Young et al. (1995), monochloramine
preferentially minimised TCA formation. Here, TCA was reduced by 92% compared to
62% for DCA.
DCA
Concentration HAs (µg/L)
10
TCA
NH2Cl - work ref. = monochloramine (e.g. NH 2Cl-B1)
8
6
4
2
BR3
Cl-B
R3
NH2
BR2
Cl-B
R2
NH2
BR1
Cl-B
R1
NH2
UR3
NH2
Cl-U
R3
UR2
NH2
Cl-U
R2
UR1
NH2
Cl-U
R1
LR
NH2
Cl-LR
L2
Cl-L2
NH2
L1
Cl-L1
NH2
B2
Cl-B
2
NH2
Cl-B
1
NH2
B1
0
Figure 5.7 Distribution of HAs after 24 hours bench scale exposure to chlorine and
117
HNMs
The final group of DBPs reported here are the HNMs (Figure 5.8). The total
concentration of HNMs measured after exposure to chlorine ranged from not detected to
3.4 µg/L (Figure 5.8). The predominant HNM was TCNM, with a concentration of up to
3.4 µg/L. This result was in line with Krasner et al. (2001) who reported TCNM
concentrations of up to 2.0 µg/L. DBNM was detected here in B1, B2, L2, LR, BR2 and
BR3, with the highest concentration found in BR2. Although other researchers have
shown that pre-ozonation can increase the formation of TCNM (Hoigné and Bader,
1988) or other HNMs (Richardson et al., 1999; Plewa et al., 2004), it was not possible
to see this trend here. The highest concentration of HNMs was observed in BR1, a
lowland river, followed by UR1, which is an upland reservoir. Hence, the nitrogenous
precursors available for the formation of TCNM are independent of the source type (e.g.
lowland, upland).
On average the concentration of HNM was reduced by 81% when using
monochloramine (Figure 5.8). Recently Hua and Reckhow (2007) found that when
using monochloramine only traces concentrations of TCNM were found and Zhang et
al. (2000) reported a decrease of 58% with monochloramine in comparison to chlorine.
TCNM
DBNM
NH2Cl - work ref. = monochloramine (e.g. NH 2Cl-B1)
3
2
1
BR3
Cl-B
R3
NH2
BR2
Cl-B
R2
NH2
BR1
Cl-B
R1
NH2
UR3
NH2
Cl-U
R3
UR2
NH2
Cl-U
R2
UR1
NH2
Cl-U
R1
LR
NH2
Cl-LR
L2
Cl-L2
NH2
L1
B2
Cl-B
2
NH2
Cl-B
1
B1
NH2
Cl-L
1
ND
0
NH2
Concentration HNMs (µg/L)
4
Figure 5.8 Distribution of HNMs after 24 hours bench scale exposure to chlorine and
monochloramine for 11 treated waters (ND – not detected)
118
5.3.3
Relationships between HAAs, THMs and other DBPs
Establishing a number of relationships between DBPs would allow a better
understanding of DBP formation and the main interest is to determine whether some
species can be used as a surrogate for others. Here the correlation between HAAs and
THMs was investigated (Figure 5.9) and it was found that for the tested waters THMs
were generally a good surrogate for HAAs when chlorine was used (coefficient of
correlation R2 = 0.82). The slope of this correlation was 1.21, which suggests that there
is slightly more than one µg of HAA formed for one µg of THM. No correlation could
be found between THM and HAAs when using monochloramine. The good correlation
between THMs and HAAs is useful for quality control and monitoring in water utilities
because, in general, laboratory analyses for HAAs cost considerably more and are also
more time consuming than the THM analyses (Sérodes et al., 2003). Nevertheless, this
correlation has its limitations. For example, a strong relationship (R2 = 0.92) has been
observed by Ates et al. (2007) while using FP tests, as well as Nissinen et al. (2002) (R2
= 0.90) in their final waters, whereas Sérodes et al., 2003 found moderate relationships
(R2 = 0.56, 0.57 and 0.63) with FP tests. Malliarou et al. (2005) reported a good
relationship between THMs and HAAs in final waters from two geographically
different regions (R2 = 0.82 and 0.90), whereas they found a poor correlation in the
waters of their third region investigated and suggested that total THM could not be
assumed to be a good indicator for HAA levels. The results here proved that THMs can
be a surrogate for HAAs, but with uniform conditions of experiments used in FP tests.
119
90
HAAs
Semi-volatile DBPs
80
70
HAAs
y = 1.21x
R2 = 0.81
60
50
40
Semi-volatile DBPs
y = 0.31x
R2 = 0.68
30
20
10
0
0
10
20
30
40
50
60
70
Figure 5.9 Correlation between THMs with HAAs and the semi-volatile DBPs (chlorine
FP tests)
This study is the first to report levels of DBPs other than just THMs and HAAs in UK
treated waters and therefore, concerns have been expressed about the practicality of
performing several DBP analyses and processing in water utility’s routine monitoring.
Previously it was shown that THMs could be used as a surrogate to determine the level
of HAAs under the uniform conditions used here. Moderate relationship was also found
between the total THM and the sum of the semi-volatile DBPs (HAN, HA, HK, i-THM
and HNM) measured after exposure to chlorine (Figure 5.9). The R2 obtained for the
collated semi-volatile DBPs was 0.68, which is in line with a previous correlation (R2 =
0.76) found between total THM and non-THM DBPs in drinking waters (Krasner et al.,
1989). This correlation suggests that the control of THM precursors is closely linked to
the control of other DBP precursors. As explained by Krasner et al. (1989) this trend is
valid for the sum of the measured halogenated DBPs but it does not give similar trends
for individual compounds; e.g. comparing total THMs to HNMs yields an R2 of only
0.08. In terms of regulation, it is interesting to note that the regulatory limit of 100 µg/L
for the THM4 would fail a regulation of 80 µg/L for the nine HAAs, currently under
consideration by the European Union (Cortvriend, 2008). Indeed from the correlation
120
found here, if 100 µg/L of THM4 would be formed, it would be expected 121 µg/L of
HAA9. To achieve a concentration of 80 µg/L for HAA9, THM4 should be concentrated
to about 65 µg/L.
Overall observations showed better correlations between species while using chlorine
than during monochloramine. No further investigation was undertaken.
5.3.4
Relationships between NPOC, UV and SUVA with DBPs
Relationships between NPOC, UV and SUVA with HAAs, THMs and the semi-volatile
DBPs were investigated with chlorine FP test data. NPOC, UV and SUVA have been
used previously as surrogates for measuring DBPs as they are easier, cheaper and faster
to measure than DBPs (Goslan et al., 2002; Parsons et al., 2005; Ates et al., 2007).
Firstly, HAAs, THMs and semi-volatile DBPs have been plotted against NPOC (Figure
5.10) and correlation (R2 values) between NPOC and HAAs, THMs and the semivolatile DBPs (collated together) were moderate (0.51, 0.63 and 0.56 respectively).
Although stronger correlations have been reported, such as White et al. (2003), who
found an R2 of 0.86 and 0.87 for HAAs and THMs respectively to NPOC, it is likely
that correlations observed in a single sample are better than correlations observed from
a range of water sources (Reckhow and Singer, 1990). It must also be considered, that a
number of other parameters, such as the bromide, iodine as well as the interactions
between parameters can affect the formation of DBPs, and therefore, the results, here,
are limited to a single analysis. In terms of semi-volatile DBPs as separate species it
was observed that NPOC correlated well with HANs (R2 = 0.82) and moderately with
HAs (R2 = 0.52) (Table 5.2). However, no correlations were found between NPOC with
i-THMs, HNMs and HKs.
121
90
HAAs
THMs
Semi-volatile DBPs
80
70
HAAs
R2 = 0.51
60
THMs
R2 = 0.63
50
40
30
20
Semi-volatile DBPs
R2 = 0.56
10
0
0
0.5
1
1.5
2
2.5
NPOC (mg/L)
3
3.5
4
Figure 5.10 Relationship between NPOC and DBPs
No correlations were found between either SUVA254 or UV254 and DBPs (Table 5.2).
For example an R2 of 0.11 was found between UV254 and HAAs, an R2 of 0.06 between
UV254 and THMs and an R2 of -0.07 between UV254 and the collated semi-volatile
DBPs. Individual groups of semi-volatile DBPs did not show any correlation with
UV254 (Table 5.2). Further correlations were investigated by considering the coagulated
waters only in relationships to DBPs. The best R2 was obtained between UV254 with the
semi-volatile DBPs (adding together) (R2 = 0.82), followed by with the HAAs (R2 =
0.78) and with the THMs (R2 = 0.49) (Figure 5.11). The investigation of the individual
groups of semi-volatile DBPs showed good relationships between UV254 with HKs (R2
= 0.72) and HAs (R2 = 0.86) and moderate correlations with i-THMs (R2 = 0.50), but no
correlations existed between UV254 with HNMs and HANs (Table 5.2).
122
80
HAAs
THMs
Semi-volatile DBPs
70
60
HAA
R = 0.78
2
50
THMs
R2 = 0.49
40
30
20
Semi-volatile DBPs
R2 = 0.82
10
0
0
1
2
3
4
UV (1/m)
5
6
7
Figure 5.11 Relationship between UV254 and DBPs for coagulated waters
Finally, the correlation between SUVA254 with DBPs was non-existent when all the data
were considered. Data presented in Table 5.2 show R2 between 0.003 and 0.27 for every
group of DBPs individually. Investigation of the coagulated waters only did not show
any improvements and it was therefore concluded that the waters investigated here did
not present any correlation with SUVA.
Table 5.2 Correlation between DBPs and water characteristics
DBPs (µg/L)
HAAs
THMs
i-THMs
HANs
HKs
HAs
HNMs
NPOC (mg/L)
0.51
0.63
0.07
0.82
0.42
0.52
0.03
Coefficient of correlation (R2)
UV (1/m)
Coagulated
All data
waters
0.11
0.78
0.06
0.49
0.09
0.50
0.09
0.45
0.11
0.72
0.11
0.86
0.03
0.06
SUVA
(L/m-mg C)
0.15
0.23
0.003
0.27
0.12
0.16
0.25
Overall, the results showed a lot of variations in correlations between DBPs and the
water characteristics. Investigation of coagulated waters only showed improved
123
correlations between UV254 and some of the DBPs, but this was not the case for NPOC
and SUVA and their correlations to DBPs. This presented work showed that in general
water characteristics could not be considered as surrogate for the measurement of DBPs
when a range of English and Welsh treated waters were collated together. Therefore,
more work is needed to investigate several samples throughout the year from one source
to determine if correlations between water characteristics and DBP formation could be
specific to individual sources as expressed by Reckhow and Singer (1990). In addition,
the results obtained here, were based on a single correlation analysis and therefore, the
results are limited.
5.4
CHAPTER SUMMARY
This study was designed to determine the formation potential of HAAs, THMs and
semi-volatile DBPs formed from a range of treated waters after exposure to chlorine or
monochloramine. General results are summarised in Table 5.3. The main results were:

Variable levels of HAAs, THMs and semi-volatile DBPs were formed.

Levels of TCM were similar to the levels of TCAA.

The prevalent classes of DBPs were HAAs and THMs; HAs were found to be
the third major group.

The impact of bromide on the formation of DBPs is well documented in the
literature and herein the data reaffirmed that more bromide species are formed in
high
bromide-containing
waters.
For
HAAs,
brominated
species
are
predominant when bromine was > 100 µg/L, whereas the brominated THMs
dominated when bromine was ≥ 75 µg/L.

The use of FP test was found unsuitable for the quantification of i-THMs.

A general decrease in all the investigated species has been observed when
shifting from chlorine to monochloramine with the one exception being 1,1DCP.

Under the controlled conditions used here, THMs correlated well with HAAs,
and the semi-volatile DBPs collated together.
124

The combination of good, moderate and non-existent correlations between water
characteristic parameters and DBP-FP were mainly attributed to the fact that
several water sources were studied, giving a complex matrix in which
correlations were not always achievable.
Table 5.3 Summary of the findings
DBPs
a
DBP levels (µg/L)
Preformed
Cl2a
NH2Clb
Shifting Cl2 to
NH2Cl
Impact of
geographical
location
Relationships
HAAs
5.0 – 68
2.1 – 14
Overall decrease
DXAA>TXAA
Variable
between sources
THMs and
NPOC
HAAs, SVeDBPs and
NPOC
THMs
2.6 – 66
0.1 – 13
Overall decrease
Variable
between sources
i-THMs NDc – 0.73
NDc – 0.89
NCd
NCd
None
HANs
0.02 – 5.5
0.02 – 0.94
Overall decrease
Highest in
lowland sources
THMs and
NPOC
HKs
0.37 – 3.9
0.01 – 1.0
1,1-DCP increase
1,1,1-TCP decrease
Variable
between sources
None
HAs
0.92 – 9.5
0.01 – 0.86
DCA slight decrease
TCA great decrease
Variable
between sources
THMs and
NPOC
HNMs
0.003 – 3.4
0.05 – 0.5
Overall decrease
Variable
between sources
None
Chlorine; b Monochloramine; c Not detected; d Not conclusive; e Semi-volatile.
125
126
Chapter 6
UNDERSTANDING AND CONTROL OF HAA AND THM
FORMATION IN TREATED WATERS
6.1
INTRODUCTION
NOM is described as an intricate mixture of organic compounds that occurs universally
in ground and surface waters. Whilst NOM itself is not problematic, it can be converted
into DBPs when in contact with disinfectants used during water treatment (Krasner et
al., 1989). In the UK, it has previously been shown (Chapter 5) that treated water has
the potential to form HAAs and THMs under specific controlled conditions.
Nevertheless, many factors influence the formation of HAAs and THMs. These factors
include the pH, the water temperature, the contact time with the disinfectant and the
disinfectant type (Liang and Singer, 2003). Furthermore, the type and concentration of
NOM as well as the concentration of bromide have a direct impact on the levels of
THMs and HAAs formed (Cowman and Singer, 1996; Hua and Reckhow, 2007).
It is known that the formation of THMs is enhanced at high pH (Carlson and Hardy,
1998), however, the effect of pH on the formation of HAAs is equivocal. Overall, HAA
formation increases with decreasing pH (Krasner, 1999). It is well established that the
DXAA concentration remains constant, while the TXAA concentration decreases with
increasing pH (Singer et al., 2002). The impact of bromide on the formation of DBPs is
well documented in the literature and the conclusions of Chapter 5 reaffirmed that more
bromide species are formed in high bromide-containing waters. In Chapter 5, it was also
found that for HAAs, brominated species are predominant when bromine was > 100
µg/L, whereas the brominated THMs dominated when bromine was ≥ 75 µg/L. The
experiments in Chapter 5 were undertaken after 24 hours contact time with the
disinfectant. Contact time is an important parameter and it is well known that while
using chlorine, DBP formation increases with increasing contact time and chlorine dose
applied (Fleischacker and Randtke, 1983). High NOM concentrations have generally
been associated with high DBP concentrations (Liang and Singer, 2003; Fearing et al.,
2004; Sharp et al., 2006). In the UK, chlorination tends to occur after the water has been
127
Chapter 6 – Understanding and control of HAA and THM formation in treated waters
treated by a coagulant and filtered (Parsons, 2006), which is different from US
treatment practices where pre-chlorination is widely used (Singer et al., 2002). After
conventional treatment processes, NOM is mainly hydrophilic in character and low in
concentration (Goslan et al., 2002). However, hydrophilic NOM has been reported to
contribute substantially to the formation of DBPs especially for waters with a low
humic (hydrophobic) content (Hua and Reckhow, 2007). Specifically, Hwang et al.
(2001) reported the hydrophilic acid + neutral fraction is the most reactive towards the
formation of THMs and HAAs, entailing the water companies to focus on the removal
of this part of NOM to have a better control of DBPs.
Water utilities are also looking towards alternative disinfectants to meet the regulations
stipulated previously in this thesis. In this direction, chloramines have been identified as
an effective alternative disinfectant because they generally result in lower
concentrations of THMs and HAAs (Speitel, 1999; Diehl et al., 2000). In Chapter 5,
HAAs and THMs have been found to decrease when shifting from chlorine to
monochloramine. With concurrent addition of chlorine and ammonia, the HAA
formation is typically 5 to 20% of that observed with chlorine alone (Speitel, 1999) and
DXAAs are the most commonly formed DBPs comprising > 90% of the total HAA
(Diehl et al., 2000).
Controlling DBPs requires extensive experimental procedures and chromatographic
analysis, which is time and cost consuming. Therefore, many models have been
developed to predict THM and HAA formation (Amy et al., 1987; Adin et al., 1991;
Chowdhury et al. 1991; Garcia-Villanova et al., 1997; Clark, 1998; Rodriguez et al.,
2000; Gang et al., 2003). However, the creation of predictive models requires a large
database of exiting results and sometimes tends to over- or under-predict DBPs (Greiner
et al., 1992). One model that shows good fitting to experimental data is the
mathematical chlorine decay model based on parallel first order reaction, detailed by
Gang et al. (2002). By integration of this model, it is possible to accurately predict
HAAs and THMs under specific experimental conditions.
The aims of this chapter were to (1) investigate the impacts of pH, temperature, bromide
concentration, contact time with chlorine and the use of alternative disinfectant on the
controlled formation of HAAs and THMs from two distinctly different waters; (2) to
128
determine the principal fractions responsible for the formation of HAAs from these two
waters and finally (3) to fit the chlorine decay data to the model proposed by Gang et al.
(2002) and to propose a model for the prediction of HAAs and THMs. The waters were
chosen from different geographical regions in the UK and collected after treatment, but
before disinfection. They have been primarily examined in the survey described in
Chapter 5, and showed potential to form HAAs and THMs. This work is specific to
THMs and HAAs, because THMs are regulated in the UK and HAAs are considered for
regulation (Cortvriend, 2008). The first and the third aim were undertaken with five
HAAs (MBAA, DCAA, TCAA, BCAA and DBAA), because of the standard
commercially available at the time of this study. Thereafter, HAA9 became available
and were hence used for the second part of this chapter. This work examines the
sensitivity of DBP formation to differences in water character and establishes whether
treated UK waters follow the trends identified for untreated US waters regarding THM
and HAA formation.
6.2
MATERIALS AND METHODS
6.2.1
Water sample and characterisation
All experiments were undertaken with samples collected in Spring 2006 from two water
utilities: Anglian Water from East Anglia (lowland water) and Yorkshire Water (upland
water). The WTWs were selected because of their different NOM content as well as
their different geographical locations. The lowland water reservoir is situated in the East
Anglian region of England in the South East. The reservoir is on a plateau with nearly
all of its water being extracted from a local river. The upland water reservoir is situated
in northern England. It is fed from a range of reservoirs set in peat-rich moorlands. A
basic description of the lowland and upland water treatments are shown in Figure 6.1
and Figure 6.2. The sampling point at each work is denoted by a *. Both samples did
not contact with any disinfectant.
129
Raw
water
PreOzonation
Distribution
Clarification
Disinfection
Intermediate
pumping
station
ASGa
Filtration
*
Final
Ozone
Carbon
contact
Chlorine
Figure 6.1 Process schematic of the lowland WTW (a Anthracite, sand and garnet)
Raw
water
Coagulation
DAFa
Sand Filtered
*
Mn
Contractors
Chlorine
Distribution
Figure 6.2 Process schematic of the upland WTW (a Dissolved air flotation)
A large volume of each water was collected (≥100 L) and stored in the fridge at 7ºC
until used. Periodic measurements of pH, NPOC, UV254 and bromide concentration
were carried out and results were consistent indicating the stability of NOM. NPOC,
UV254, and pH analytical measurements are as described in Chapter 5, Section 5.2.
6.2.1.1 Measurement of bromide using IC
Analysis of bromide was carried out with an IC system, (Dionex, DX500 series, UK).
This consisted of an IonPac AG9HC guard column (4 x 50 mm), an IonPac AS9HC
separation column (4 x 250 mm), an anion electrolytic suppressor (Atlas, 4 mm, UK) in
auto-suppression recycle mode, an SC 20 suppressor controller, an ED 40 in
conductivity mode and a 250 μL sample loop (all supplied by Dionex, UK).
Instrumental control and data collection were performed using PeakNet 5.11
chromatography workstation (Dionex, UK). All the tubing in the chromatography path
(from pump outlet to exit of cell) was PEEK (0.125 mm ID). The eluent was 9 mM
130
sodium carbonate (Na2CO3) anhydrous (Fisher Scientific, UK). The suppressor current
was set to 72 mA and the eluent flow rate was kept at 1.0 ml/min. A calibration curve
was made with standards of bromide at 10, 25, 50, 75, 100 and 200 µg/L.
6.2.1.2 Alkalinity
Alkalinity is the acid neutralising capacity of the water. It can be determined by
titration. Alkalinity depends on the end point used in the titration. For routine analysis,
an end point of pH 4.5 is used. Titration was made with 0.02 M hydrogen chloride
(HCl). Bromocresol green/methyl red indicator was prepared as specified in the
Standard Method 2320B (APHA, 1998).
A sufficiently large sample to provide a good volumetric precision but small enough to
give a good end point was used. The standard method recommends a volume titrant of
about 20 mL. The volume of sample was thus placed in a conical flask. Pure water was
added to make-up the volume to approximately 100 mL. Two to three drops of
bromocresol green/methyl red indicator were then added to the volume of sample. Then
the titration was carried out with 0.02 M hydrogen chloride until the end point was
reached (using the bromocresol green/methyl red indicator gives a colour change from
blue to pale yellow). Then the alkalinity is calculated using the following equation:
Alkalinity, mg CaCO3/L =
At  N  50,000
mL sample
Equation 6.1
where At = mL standard acid titrant and N = normality of standard acid used.
6.2.2
Fractionation
To determine the hydrophilic/hydrophobic NOM ratio, 50 litres of the treated waters
were fractionated by XAD and cation exchange resin adsorption techniques into their
hydrophobic neutral, hydrophobic acid, transphilic, hydrophilic base and hydrophilic
acid + neutral fractions. The method used was adapted from Leenheer et al. (2004). The
column capacity factor (k’) was 60 for both samples (Goslan, 2003). The recovery of
NOM was 90 and 113% for the lowland and upland water respectively.
131
The resins used were Amberlite XAD-7HP resin and Amberlite XAD-4 resin (Rohm &
Haas, Germany). Amberlite XAD-7HP is an acrylic ester polymer and is equivalent to
XAD-8; Amberlite XAD-4 is a styrene DVB polymer. Amberlite 200 strongly acidic
cation exchanger has a sulfonated polystyrene/DVB matrix (Sigma-Aldrich, UK). The
XAD resins were precleaned by sequentially Soxhlet extracting for 48 hours each with
methanol, acetonitrile and methanol again to remove impurities. Before use the resins
were packed into columns and rinsed with pure water until the column effluent NPOC
was < 2 mg/L (Malcolm and MacCarthy, 1992).
The fractionation procedure is presented below in Figure 6.3.
CH3CN:H2O
(75:25)
XAD-7HP
HPO-N
Acidification to pH 2
50 L of
treated
water
CH3CN:H2O
(75:25)
XAD-7HP
HPO-A
CH3CN:H2O
(75:25)
XAD-4
TPI
NH4OH
3M
Cation
Exchange
resins
HPI-B
HPI-A+N
Figure 6.3 Schematic of the fractionation procedure
Fifty litres of lowland water was passed through XAD-7HP resin. The column was back
eluted with 400 mL of a solution of acetonitrile (CH3CN:H2O (75:25)). The
hydrophobic neutral (HPO-N) was desorbed from the resin. The effluent collected was
acidified to pH 2 and again put through the XAD-7HP column. Then the column was
back eluted with 400 mL of the same solution (CH3CN:H2O (75:25)). The hydrophobic
acid (HPO-A) was desorbed from the resin. The eluate was put through the XAD-4 and
the transphilic fraction (TPI) was desorbed by back eluting the resin with again 400 mL
132
of CH3CN:H2O (75:25). Finally the effluent was passed through the cation exchange
resin. The overall effluent was the hydrophilic acid + neutral fraction (HPI-A+N),
whereas the desorbed fraction (with 3M of ammonium hydroxide (NH4OH)) was the
hydrophilic base fraction (HPI-B).
After cleaning the XAD resins with pure water and regenerating the cation exchange
resins (using three to six bed volume of 1 to 10% of acid sulphuric as a regenerant, with
a flow between 1 to 10 m/h), 50 litres of the upland water was fractionated using the
same procedure.
6.2.3
DBP-FP
6.2.3.1 Sample preparation for the determination of parameter impacts and the
modelling study
Preparation of the solutions and details of the suppliers are given in Chapter 5, Section
5.2.3. Here, samples were chlorinated at pH 6, 7 and 8 at room temperature (20ºC) to
determine their DBP-FP. Phosphate buffer was used to adjust the pH. In addition one
set of samples for each water was chlorinated at pH 7 with addition of bromide (200
µg/L) and another set at pH 7 with a temperature of 7°C. Hypochlorite solution was
prepared using the Standard Method 4500-Cl B (APHA, 1992). (Chapter 5, Section
5.2.3). The chlorine dose required was determined by preliminary chlorine demand
experiments in order to have the free chlorine residual slightly higher than 1 mg/L as
Cl2 after seven days of contact time. A 100 mL bottle was partly filled with the water
sample, the buffer and the chlorine solution were added and the bottle was filled
completely and capped headspace free with a PTFE-lined cap. Samples were incubated
for 0.5, 1, 3, 6, 24, 72 and 168 hours at 20°C in the dark with the exception of the
samples incubated at 7°C. At the end of the incubation period the chlorine residual was
measured using Iodometric Method I (APHA, 1992) and a sulphur-reducing agent
(sodium sulphite) at a concentration of 100 g/L was added to the samples to destroy the
chlorine residual whilst not degrading the five HAAs measured (MBAA, DCAA,
TCAA, BCAA and DBAA) (Singer et al., 2002). Samples were prepared in duplicate
independently.
133
Chloraminated samples were prepared with two different ratios of preformed
chloramines: Cl2:N: 3:1 and 7:1. Two different pHs were investigated: pH 6 and 8, both
at 20°C for 168 hours. Chloramine residual was measured using the method 4500-Cl F.
DPD Ferrous Titrimetric Method (APHA, 1992). Samples were prepared in duplicate
independently.
6.2.3.2 Quality control samples
5 mg/L chlorine dosing solution was put into 100 mL glass bottle PTFE-lined screw cap
with pure water. Appropriate amounts of buffer to adjust the pH to 6, 7 and 8 were
added. Then the bottle was filled with pure water and stored with the samples at 20 ºC
for up to seven days.
6.2.3.3 Sample preparation for the determination of HAA precursors
Fractions were first dried to remove any acetonitrile left in the samples. Then they were
regenerated with a solution of 0.1 M sodium hydroxide (NaOH) and their NPOC was
measured. Then they were chlorinated in a 100 mL bottle. pH was adjusted in each
fraction to 7 ± 0.2 to have comparable results. Fractions were chlorinated with 5 mg/L
of sodium hypochlorite for 72 hours and at 20ºC. At the end of the incubation period the
chlorine residual was measured using Iodometric Method I (APHA, 1992) and
ammonium chloride at a concentration of 100 mg/L was added to the samples to destroy
the chlorine residual whilst not degrading the nine HAAs measured in the fractions
(Singer et al., 2002). Samples were prepared in duplicate independently.
6.2.3.4 HAA and THM sample preparation
For the measurement of THMs, 5 mL of water sample was transferred into a 10 mL vial
allowing 5 mL of headspace. Following this, the samples were analysed by headspace
GC/MS. Samples were prepared in duplicate and analysed in triplicate. HAA samples
were first converted to their protonated forms before processing the extraction with
organic solvent and deriving to form methyl esters. The method used for the
derivatisation was extensively explained in Chapter 3.
134
For both HAAs and THMs, water samples were extracted without being in contact with
disinfectant to determine the natural and anthropogenic presence of THM and/or HAAs.
6.2.4
Sample analytical methods
HAA5 measured for the determination of the parameter impact and the modelling study
were quantified with an Agilent 6890 GC/ECD available at Open University (Milton
Keynes, UK). A volume of 1 µL was injected with the injector at 200°C with a 5:1 split,
separation was performed by a BPX5 column (SGE, UK; 30 m × 0.25 mm × 0.25 µm)
with a helium carrier gas at a column flow rate of 1.1 ml/min. The initial oven
temperature was 35°C followed by a 5°C per minute temperature ramp to 220°C and
held for 1 minute. The detector temperature was 230°C and the rate of data collection
20 Hz. This was carried out at the Open University (Milton Keynes, UK).
Instrument conditions for the measurement of HAA9 from the fractionated samples were
described in Chapter 3 Section 3.3.1.4.
THMs were analysed using a Varian Saturn 2200 (ion-trap) GC/MS. The samples were
heated and agitated by CTC CombiPal to 60°C for 30 minutes. 500 µL of headspace
was removed by heated syringe and injected with a 10:1 split, separation was performed
by a BPX5 column (SGE, UK; 30 m × 0.25 mm × 0.25 µm) with a helium carrier gas at
a column flow rate of 1.1 ml/min. The injector temperature was 250°C; the initial oven
temperature was 45°C for 2 minutes followed by a 10°C per minute temperature ramp
to 90°C. The MS was operated in the electron ionisation (EI) mode. The ion-trap
temperature was set at 230°C and the electron energy was 70 eV. Mass spectra were
collected in full scan mode (33-300 amu). The ions of 83, 129 and 173 m/z were
selected as quantification ions. Quantification of THMs was achieved by comparing the
chromatograms of the samples with the calibration curves from standards. THM
measurement was carried out at the Open University (Milton Keynes, UK).
135
6.3
RESULTS AND DISCUSSION
6.3.1
Water characterisation
Characteristics of the waters are summarised in Table 6.1. The concentration of NOM
was greater in the lowland (4.7 mg/L) than in the upland water (2.1 mg/L). NOM
fractionation indicated that NOM from the upland water had a higher hydrophilic
content than the lowland water which had significant transphilic content (Table 6.1).
Hwang et al. (2001) reported that the transphilic fraction of intermediate polarity is
generally more hydrophobic than hydrophilic but this statement was highly dependent
on the water source.
The reactivity with respect to DBP-FP can be characterised with SUVA254 (Edzwald
and Tobiason, 1999). A high SUVA254 value is an indicator of a high reactivity towards
DBP production (Singer et al., 2002). The lowland water had a lower SUVA254 value
than the upland water (1.3 and 2.3 L/m-mg C respectively), but both waters had overall
relatively low SUVA254 values as compared to other fresh waters reported in the
literature (Edzwald and Tobiason, 1999). In addition, these values were similar to the
values found in the same water and studied in Chapter 5 (SUVA254 values of 1.6 and 2.6
L/m-mg C for lowland and upland water respectively).
Table 6.1 Water characteristics
Parameters
pH
NPOC (mg/L)
UV254 (1/m)
SUVA254 (L/m-mg C)
Alkalinity (mg/L of CaCO3)
Bromide content (µg/L)
THM-FPa (µg/L)
HAA-FPa (µg/L)
Hydrophobic neutral (%)
Hydrophobic acid (%)
Transphilic (%)
Hydrophilic base (%)
Hydrophilic acid + neutral (%)
Upland water
6.7
2.1
4.8
2.3
6
34
72
104
4
19
8
2
67
Lowland water
8.0
4.7
5.9
1.3
188
206
89
84
2
23
31
4
40
a
Experimental conditions: pH = 7, temperature = 20°C and contact time with chlorine
= 168 hours.
As expected, the two waters differed not only in their alkalinity but also in their
bromide concentration. The bromide concentration of the lowland water (206 µg/L) was
136
six times higher than that of the upland water (34 µg/L). It is therefore likely that the
lowland water will produce more brominated species.
HAA5 (MBAA, DCAA, TCAA, BCAA, DBAA) and THM4 concentrations were similar
in the lowland water after 168 hours contact time at pH 7, whereas the upland water had
the potential to form more HAA5 than THM4. Same trend was observed in Chapter 5 for
the nine HAAs after 24 hours exposure to chlorine. It should be considered that the
quantification of the three remaining HAAs (not measured here) might contribute to
higher concentration of the total HAA in the lowland water considering the high level of
bromide. For example in Chapter 5, it was observed that after 24 hours contact time
with chlorine, HAA3 (BDCAA, DBCAA and TBAA) contributes to 27% of the total
HAA in the lowland water, but only 7% in the upland water. Malliarou et al. (2005)
found that some regions of the UK produced an average total level of THMs higher than
the HAAs, while the contrary was found in other regions. This highlights the differences
observed in different geographical locations in the UK.
6.3.2
Parameters affecting the formation of HAAs and THMs
6.3.2.1 Impact of pH
It is well known that the formation of DBPs is strongly dependent on the pH set while
using chlorine as to disinfect (Krasner, 1999; Singer, 1999; Xie, 2003). Here it has been
shown (Figure 6.4) that increasing pH from 6 to 8 decreased HAA5 in the lowland water
by 15%. DCAA and BCAA were found to be more affected by pH than TCAA. DCAA
and BCAA concentrations at pH 8 were significantly lower than at pH 6 and 7. Liang
and Singer (2003) reported that increasing pH from 6 to 8 had a very little effect on the
formation of the MXAA and DXAA species, but significantly decreased the formation
of the TXAA species. In the literature, DCAA formation was reported to be highest at
pH 7 (Krasner, 1999) which is true of the results found here.
137
100
90
80
pH 6 Contact time = 168 hours
pH 7
pH 8
70
60
50
40
30
20
10
0
MBAA
DCAA
TCAA
BCAA
DBAA
HAA5
HAA species
Figure 6.4 Comparison of pH effect on the formation of measured HAA5 in the lowland
water
For the upland water HAA5 formation was 14% greater at pH 6 than at pH 8 and the
lowest concentration was found at pH 7 (Figure 6.5). In the upland water, only TCAA,
DCAA and BCAA were detected due to the low bromide content in this water. The
DCAA concentration was higher at pH 6 and similar at pH 7 and 8. TCAA increased by
28% with increasing pH, which is contrary to the literature (Liang and Singer, 2003;
Zhuo et al., 2001) and the results found for the lowland water.
The contradictory results found here were believed to be a result of the difference of
NOM present in the waters. Although Singer et al. (2002) concluded that precursors for
MXAA, DXAA and TXAA were different, here it can be concluded that the precursors
of DXAA and TXAA did not react with chlorine in a similar way from one source to
another. This would suggest that the precursors between sources could have the same
base chemical structure, but may have different functional groups attached to this base
and therefore the pH did not promote the oxidative cleavage of the functional group the
same way between sources, which could induced the contradictory results observed
here.
138
160
140
120
pH 6
pH 7
pH 8
Contact time = 168 hours
100
80
60
40
20
0
MBAA
DCAA
TCAA
BCAA
DBAA
HAA5
HAA species
Figure 6.5 Comparison of pH effect on the formation of measured HAA5 in the upland
water
Trussell and Umphres (1978) reported that the formation of THMs consists of alternate
hydrolysis and halogenation steps. All these reactions are favoured under alkaline
condition, thus more THMs are formed at higher pH, which is illustrated by the results
found here after 168 hours exposure to chlorine (Figure 6.6). The impact of pH is
limited in the upland water (not shown) compared to the lowland water which could be
explained, again, by the difference of NOM responsible for the THM formation and its
likelihood to undergo hydrolysis and halogenation reactions.
139
140
120
pH 6
pH 7
pH 8
100
80
60
40
20
0
TCM
BDCM
DBCM
TBM
THM4
THM species
Figure 6.6 Comparison of pH effect on the formation of THM4 in the lowland water
Because the impact of pH is equivocal on the formation of HAAs, but well defined on
the formation of THMs, it cannot be concluded here if there is a link between the HAA
and THM precursors. The impact of pH combined with the impact of contact time will
be modelled in the Section 6.3.4.
6.3.2.2 Impact of bromide
The effect of bromide concentration on HAA and THM formation and speciation was
investigated by spiking the lowland and the upland water with 200 µg/L of bromide. In
the lowland water, the addition of bromide had a slight impact (10% decrease) on the
total concentration of HAA measured. Less DCAA and TCAA were formed, whereas
more brominated HAA species were produced as expected. The addition of bromide had
a greater impact in the upland water than in the lowland water. With the upland water
the concentration of HAA5 decreased (Figure 6.7) with a switch from DCAA and
TCAA to brominated species MBAA, BCAA and DBAA.
140
It was reported by Hua et al. (2006) that the total concentration of the five regulated
HAAs in the US (MCAA, MBAA, DCAA, TCAA and DBAA) decreased as bromide
concentration increased because of the number of brominated species measured. This
applies here but the exception is that BCAA is included in the total HAA and not
MCAA. However the same study reported that addition of bromide increased the total
HAA9 yield between 0 and 35%.
120
100
[Br-] in source
Addition of 200 µg/L [Br-]
Contact time = 168 hours, pH = 7
80
60
40
20
0
MBAA
DCAA
TCAA
BCAA
HAA species
DBAA
HAA5
Figure 6.7 Impact of bromide on the formation of measured HAA5 in the upland water
Bromine was reported by Cowman and Singer (1996) to be more reactive than chlorine
in substitution and addition reactions that form HAAs, thus the inclusion of bromine
shifts the speciation of the HAA towards the brominated species. Here, the difference
observed in bromide incorporation is likely to be due to the number of HAA species
measured.
The formation of THM is also affected by the addition of bromide. Hua et al. (2006)
reported that increasing initial bromide levels resulted in a substantially increased THM
molar concentration between 14% and 74%. Here the total THM weight concentration
increased by 60% in the lowland water and by 54% in the upland water. In the upland
141
water, only the brominated species increased, whereas all the brominated species and
TCM were augmented in the lowland water.
6.3.2.3 Impact of temperature
Reducing the incubation temperature from 20°C to 7°C, resulted in a reduction of both
HAAs and THMs. The concentration of HAAs and THMs dropped by 59% and 43%
respectively in the lowland water (Figure 6.8) and by 43% and 53% respectively in the
upland water.
120
Lowland water
Upland water
100
80
60
40
20
0
HAA @ 20°C
HAA @ 7°C
THM @ 20°C
THM @ 7°C
Figure 6.8 Temperature effect on the DBP-FP in lowland and upland water
El-Dib and Ali (1995) reported that the effect of temperature (rise between 0 and 30°C)
on the THM yield was rather limited compared with data reported by other investigators
(Urano and Takemasa, 1986) and concluded that the differences were due to the nature
of organic precursors liable to be found in the water.
Dojlido et al. (1999) reported that the concentrations of HAAs were seasonally
dependant. During the winter season (1°C) they found levels of ~ 0.63 µg/mg C
whereas in the summer (23°C), concentrations reached ~ 7.4 µg/mg C. In the UK, the
142
effect of season on HAA formation has not been determined but the results shown here
(Figure 6.8) indicate there may be a seasonal effect.
6.3.2.4 Impact of reaction time
The rate of DBP formation (Figure 6.9) was characterised by a general trend of an initial
rapid reaction during the first few hours, followed by a steady rate of increase over the
168 hours of contact time with chlorine, and this agrees with Reckhow et al. (1990) and
Singer et al. (2002). After six hours of reaction, up to 30 and 40% of the reaction has
been completed for the lowland and upland water respectively, increasing to 50 and
65% after 24 hours, again for the lowland and upland water respectively. Singer et al.
(2002) reported that as long as NOM and free chlorine were present, both HAAs and
THMs would continue to form. The µg levels of HAA and THM formed per mg of
NPOC were respectively 2.8 and 1.8 greater in the upland water than in the lowland
water after 168 hours contact time with chlorine, despite the fact that the NPOC of the
upland water was 2.1 mg/L and the one of the lowland water was 4.7 mg/L (Figure 6.9).
This is confirmed by Peters et al. (1991), who reported that water from a surface river
with a DOC value of 4.6 mg/L produced 3.8 µg/L total HAA, whereas they found other
surface water with a DOC of 2.7 mg/L forming 10 µg/L total HAA, highlighting the
water specificity of NOM reactivity. Presently, the NOM available in the treated
lowland water showed similar HAA5 and THM4 outcome when in contact with chlorine,
whereas in the upland water, the HAA5 formed faster at the beginning of the reaction
than the THM4. For example after six hours of contact time with chlorine the upland
water exhibited a HAA5 concentration (20 µg/mg C) 1.7 greater than the THM4
concentration (12 µg/mg C). It could be deduced that the reactive site of NOM from
each water differed. In addition, the difference observed between the two waters is also
believed to be due to the number of HAAs measured. Here, five HAAs are quantified;
however, lowland water has high level of bromide, which can account for more
brominated species, not quantified here. Furthermore, the result obtained here for the
HAA5 after 24 hours contact time with chlorine tallied with the results found in Chapter
5 for the five HAAs. Therefore, the reactivity of NOM remains similar between water
collected in 2006 and 2008.
143
[DBP]/NPOC (µg/mg C)
60
LW - HAA
LW - THM
UP - HAA
UP - THM
50
40
pH = 7
30
20
10
0
0
20
40
60
80
100
120
Contact time (hours)
140
160
180
Figure 6.9 Impact of chlorine contact time on HAA5 and THM4 formation from lowland
water (LW) and upland water (UW), experimental conditions: pH = 7°C, temperature =
20°C, NPOC of lowland water = 4.7 mg/L and upland water = 2.1 mg/L
6.3.2.5 Impact of disinfectant
The impact of switching from chlorine to chloramines in both waters is presented in
Figure 6.10 and Figure 6.11. As shown, pH 6, pH 8 and two chlorine to nitrogen ratios
(Cl2:N) of 3:1 and 7:1 for the preformed chloramines have been investigated after 168
hours contact time with the disinfectant. Firstly, there was a considerable decrease of
total HAA in both waters when using chloramines as an alternative to chlorine. In the
lowland water, total HAA formation decreased between 82 to 92% (Figure 6.10), which
is similar to the decrease observed in the upland water between 86 to 94% (Figure 6.11).
These values compare well with Cowman and Singer (1996), who reported a reduction
on the order of 90 to 95%, when using chloramines instead of chlorine. In any case,
here, waters exposed to preformed chloramines did not form more than 20 µg/L total
HAA, whereas it was observed the lowland and the upland water exposed to chlorine
could form up to 80 and 133 µg/L of HAA5 respectively. In the lowland water (Figure
6.10), the major decrease was obtained at pH 6 and with the Cl2:N ratio 3:1, whereas in
the upland water (Figure 6.11), this was at pH 8, with the same Cl2:N ratio of 3:1. Diehl
144
et al. (2000) reported that HAA concentration in general decreased as the pH increased
(here, pH 8) and the Cl2:N ratio decreased (here, ratio 3:1). Here the results agree for the
upland water, whilst the contrary to Diehl et al. (2000) was found for the impact of pH
in the lowland water. The impact of pH was reported to be equivocal in Section 6.3.2.1
and can explain the difference observed between waters and with Diehl et al. (2000).
However, the chloramine chemistry is well known. A Cl2:N ratio of 7:1 is close to the
breakpoint chlorination and, therefore, the species present are a mixture of
monochloramine, dichloramine and free chlorine. It is the free chlorine, which can be
responsible for more HAA formed at the Cl2:N ratio of 7:1, whereas at a Cl2:N ratio of
3:1, only monochloramine is present (Wolfe et al., 1984) and the lack of free chlorine is
the reason for the low formation of HAAs, as it has been pointed out by Diehl et al.
(2000).
90
80
70
60
DBAA
BCAA
TCAA
DCAA
MBAA
50
40
pH 8
pH 6
30
20
10
0
3:1
7:1
Chloramines
Chlorine
3:1
7:1
Chloramines
Chlorine
Figure 6.10 Comparison of chlorine versus chloramines in lowland water, at pH 6 and
pH 8; Cl2:N ratio = 3:1 and 7:1, contact time = 168 hours
The major species detected with chloramines were the DXAAs in both waters, with a
contribution between 86 to 95% of the total HAA at both pHs, both ratios. This is in line
with Diehl et al. (2000), who also reported DXAA to be the dominant HAA species,
145
comprising between 70 to 100% of the total HAA6 (HAA5 + MCAA). The difference
observed between the two waters, here, consists of the speciation of the DXAAs. In
Chapter 5, Section 5.3.2.1, the BIF was introduced and it characterises the degree of
bromine substitution. Here, the lowland and the upland water had 206 and 34 µg/L of
bromide respectively. Therefore after exposure to chloramines, BIFs were significantly
different from lowland to upland water. For example, BIF values ranged from 0.53 to
0.99 for the lowland water, whereas they ranged from 0.03 to 0.16 in the upland water.
The values of BIF compare well with Pope et al. (2006), who also reported higher BIF
into DXAA formed with chloramines, when the water source had higher level of
bromide. For example, they reported a BIF of 0.5 for a bromide level of 168 µg/L, and a
BIF of 0.06 for a water with a level of bromide of 33 µg/L. Therefore, this explained
here the greatest formation of brominated DXAA (BCAA and DBAA) (48 to 68%) in
the lowland water than in the upland water (3 to 18%). Consequently in the upland
water, the major species was DCAA comprising more than 90% of the species formed
with chloramines.
140
120
100
DBAA
BCAA
TCAA
DCAA
MBAA
80
60
pH 6
pH 8
40
20
0
3:1
7:1
Chloramines
Chlorine
3:1
7:1
Chloramines
Chlorine
Figure 6.11 Comparison of chlorine versus chloramines in upland water, at pH 6 and pH
8; Cl2:N ratio = 3:1 and 7:1, contact time = 168 hours
146
The effect of using chloramines instead of chlorine was also investigated on the
formation of THM4 after 168 hours contact time with the disinfectant. The results from
Table 6.2 showed that THM concentrations decreased between 75 and 98% in the
lowland water and between 92 to 95% in the upland water. The concentrations of THMs
measured in the lowland water were lower than 15 µg/L at any pHs and ratios and were
lower than 3 µg/L in the upland water. Chloramines limit the free chlorine contact time
and therefore induce a considerable reduction of THM as explained by Carlson and
Hardy (1998).
Table 6.2 Decrease of THM4 in chloraminated lowland and upland water compared to
the same chlorinated sample; contact time with chloramines = 168 hours
Cl2:N ratio
3:1
7:1
6.3.3
Decrease of THM4 in chloraminated water vs. chlorinated water (%)
Lowland water
Upland water
pH 6
pH 8
pH 6
pH 8
88
98
95
98
75
88
92
98
HAA precursor investigation
To aid the water companies to better control the DBPs, it was important to determine
their precursors in order to develop possible options for the precursor removal. While
all the previous work focused on five HAAs, it is important to specify that here the nine
HAAs have been measured, because HAA9 standards were simply available as a
mixture at the time of the experiments.
As explained earlier, the waters fractionated here were originally treated and were
therefore more hydrophilic in content. The fraction mass distribution of NPOC was
again presented in Table 6.3, and was explained earlier in Section 6.3.1. Results
presented in Table 6.3 showed that the individual NPOC fraction values were relatively
low, when doing the 72 hours FP test with chlorine. Moreover, HAA concentrations
were standardised with the NPOC values to have the best comparison between fractions.
Each fraction was evaluated for its chlorine demand. Results available for the lowland
water (Table 6.3) showed that it was the transphilic fraction that was the most active in
consuming chlorine, whereas the hydrophobic acid and hydrophilic acid + neutral
fractions consumed similar quantity of chlorine (1.3 mg/mg C). In the upland water, the
transphilic and the hydrophobic neutral fraction were the most active fractions towards
147
the chlorine demand (2.6 mg/mg C), followed by the hydrophobic acid (1.9 mg/mg C)
and the hydrophobic base equivalent to the hydrophilic acid + neutral (0.8 mg/mg C).
This differs with the results reported by Hwang et al. (2001) in which the order of
reactivity of the NOM fractions towards the chlorine demand was hydrophilic base >>
hydrophilic acid + neutral > transphilic ~ hydrophobic. However, in the literature,
Croué et al. (2000) reported that in filtered waters, the transphilic fraction was the most
active in consuming chlorine and that all the other fractions had a chlorine demand
values of 1 ± 0.2 mg/mg C, which showed that these values were water specific,
because the precursors differ from one source to another and also explained why the
chlorine demand was not similar between the lowland and the upland water. Analysis of
molecular weight showed similar trend between the two waters (Table 6.3), which
indicates that the molecular weight of the fractions was not related to their reactivity
with chlorine.
Table 6.3 Summary of NOM fraction properties in lowland and upland water
Lowland water
Fraction mass distribution of
NPOC (%)
NPOC fraction a (mg/L)
Chlorine demand (mg/mg C)
DXAA/TXAA ratio
Molecular weight b
Upland water
Fraction mass distribution of
NPOC (%)
NPOC fraction a (mg/L)
Chlorine demand (mg/mg C)
DXAA/TXAA ratio
Molecular weight b
HPO-N
HPO-A
TPI
HPI-B
HPI-A+N
2
23
31
4
40
1.9
NA c
1.5
4
0.9
2.6
0.9
1.0
1.7
1.0
1.3
2.6
NA c
1.1
2.6
1.0
HPO-A ~ TPI > HPO-N ~ HPI-B
19
8
2
1.2
1.4
0.5
1.9
2.6
0.8
0.6
0.9
1.0
HPO-A > TPI > HPO-N ~ HPI-B
1.4
1.3
2.2
NA c
67
2.5
0.8
1.7
NA c
a
Corresponds to the NPOC values of the fractions when doing experimental analyses; b Measured with
high performance size exclusion chromatography (HPSEC) (Shimadzu, Milton Keynes, UK); c Not
analysed.
Nine HAAs were measured from each fraction and the concentrations in µg/mg C were
presented in Figure 6.12. Concentrations of HAA9 ranged from 14 to 41 µg/mg C in the
lowland water and from 7 to 72 µg/mg C in the upland water. The order of reactivity of
the five NOM fractions with respect to formation of HAA9 was as follow:
 Lowland water: HPI-B > TPI > HPI-A + N > HPO-A > HPO-N
148
 Upland water: HPO-A > HPO-N > TPI > HPI-A + N > HPI-B
Note: HPO-N = hydrophobic neutral; HPO-A = hydrophobic acid; TPI = transphilic; HPI-B =
hydrophilic base and HPI-A +N = hydrophilic acid + neutral.
These results indicated a major difference of precursor reactivity towards the HAAs
between the lowland and the upland water. Hwang et al. (2001) concluded that the
reactivity of the hydrophilic base could be variable depending on the water source, but
found a general order of fraction reactivity being hydrophilic acid + neutral > transphilic
~ hydrophobic. Marhaba and Van (2000) reported that the most reactive fraction was
the hydrophobic neutral in filtered water. Therefore, fraction data varied considerably
between sources.
HAA speciation was also studied from each fraction. It was clear from the Figure 6.12
that most of the bromide was present in the last two fractions (hydrophilic base and
hydrophilic acid + neutral). The ratio DXAA:TXAA (also reported in Table 6.3)
showed a predominance of DXAA species in the lowland water, whereas this was
variable in the upland water. For example, the ratio was 0.6 for the hydrophobic acid in
the upland water, indicating a dominance of TXAA, and especially TCAA, whereas this
ratio was 1.7 in the hydrophilic acid + neutral fraction, but this can be explained by the
presence of bromide in the fractions, which is responsible for the decrease of TCAA
(explained in the Chapter 5, Section 5.3.2.1). The predominance of DXAA in the
lowland water and specifically in the transphilic and hydrophilic acid + neutral is
believed to be caused by the presence of algae in the water source. Indeed, proteins and
polysaccharides, present in these two fractions (Hwang et al., 2001), are the main
component of extracellular organic matter (EOM) of Anabaena flos-aquae, a common
blue-green algae found in reservoirs (Huang et al., 2008). Huang et al. (2008) reported
that EOM formed predominantly DXAA, when the algae were in their death phase.
No trend could be observed between the ratio DXAA:TXAA. Whilst, it was found that
DXAA correlated well with TXAA in the upland water (R2 = 0.93), no correlation could
be found in the lowland water. Hwang et al. (2001) found the order of DXAA:TXAA
ratio to be transphilic > hydrophobic and that the DXAA:TXAA ratio was variable
between water source, which tends to be similar to the results found here.
149
HAA concentration (µg/mg C)
80
70
TBAA
DBCAA
BDCAA
60
DBAA
TCAA
50
BCAA
DCAA
40
MBAA
MCAA
30
20
10
0
LW
-N
PO
-H
W
U
-H
-N
PO
LW
-A
PO
-H
W
U
-A
PO
-H
LW
PI
-T
W
U
-T
Fractions
PI
LW
-B
PI
H
-
UW
-B
PI
H
-
+N
+N
-A
-A
I
I
P
P
-H
-H
W
LW
U
Figure 6.12 HAA concentrations from lowland water (LW) and upland water (UW)
from five different fractions; experimental conditions: pH = 7, contact time = 72 hours,
temperature = 20°C
The results found previously are based on the level of HAAs formed per mg of carbon,
however, each fraction contributed to a different percentage of the mass distribution of
NPOC and therefore, it was necessary to consider the real percentage of the HAA9,
which is based on the concentration of HAA9 in µg/mg C multiply by the contribution
of the specific fraction to the fraction mass distribution of the total NPOC. Results are
presented in Figure 6.13 and showed that the fractions that contribute the most to the
formation of HAA9 in the lowland and upland water were the hydrophilic acid + neutral
(40%) and the hydrophobic acid (37%) respectively. Also in the lowland and upland
water, the transphilic also contributed 36% and 28% of the formation of HAA9
respectively. These results are based on one sample from each source; therefore, more
work should be done to confirm the results found here in order to help the water
companies tackle the right fractions when considering precursor removal.
150
Proportion HAA in fraction reconstituted (%)
50%
45%
HPO-N
HPO-A
40%
TPI
35%
HPI-B
30%
HPI-A+N
25%
20%
15%
10%
5%
0%
Lowland water
Upland water
Figure 6.13 Fractions to be targeted for their removal
6.3.4
Modelling of the chlorine decay and prediction of HAA5 and THM4
As a result of the complex nature of DBP precursor compounds and their corresponding
reactions with disinfectants, models for the quantification of DBPs have been largely
developed using empirical approaches (Westerhoff et al., 2000). In the present study,
chlorine decay over time has been measured for the lowland and the upland water at
three different pHs. Gang et al. (2002) proposed an empirical mathematical model of
chlorine decay, from which it is possible to predict THMs and HAAs, but under the
specific conditions of pH 8 and temperature of 25ºC. Therefore, it was interesting to see
if this model could first, fit the present data and secondly, to test if the model was
appropriate outside the experimental conditions specified above.
6.3.4.1 Chlorine decay data modelling
The mathematical expression development and the assumptions on which the model is
based are reported in a study of Gang et al. (2002). To summarise, it was assumed that
the formation of DBPs results in two parallel reactions, one with a rapid rate of chlorine
decay and the other a slower, long term chlorine consumption. Others assumptions were
151
that the rate of reaction for each chlorine-consuming reaction was first order in chlorine
concentration, that the chlorine demand was solely attributed to the reaction with NOM
and proportional to the formation of DBPs.
The parallel first order reaction model used to evaluate the chlorine decay was the
following:


Ct   C 0 fe  k R t  (1  f)e  k S t ,
Equation 6.2
where C(t) is the chlorine concentration at any time t (mg/L); C0 the initial chlorine
concentration dose to give the chlorine residual of at least 1 mg/L after seven days of
reaction; f the fraction of the chlorine attributed to the rapid reaction; kR the first order
rate constant for the rapid reaction (h-1); kS the first order rate constant for the slow
reaction (h-1). The parameters f, kR and kS were determined with non-linear regression
analysis using the software SPSS version 17.
Data from Figure 6.14 A+B showed that the chlorine decay was rapid during the first
six hours of the reaction followed by a more gradual decay, and this at pH 6, 7 and 8.
Attempts were then made to fit the chlorine decay data to the parallel first order reaction
model. Lines in Figure 6.14 A+B, representing the models, showed that the model fitted
the data, generating R2 between 0.956 and 0.997 (Table 6.4). Therefore, the model
proposed by Gang et al. (2002), can be adjusted to any type of waters (lowland, upland)
and at any pHs.
8
LW - pH 6
LW - pH 7
8
A
LW - pH 8
UW - pH 6
UW - pH 7
B
UW - pH 8
7
6
Chlorine decay (mg/L)
Chlorine decay (mg/L)
7
5
4
3
2
1
6
5
4
3
2
1
Models
Models
0
0
0
24
48
72
96
120
144
168
0
24
48
72
96
120
144
Figure 6.14 Chlorine decay data of (A) lowland water (LW) and (B) upland water
(UW), both at pH 6, 7 and 8 fitted to parallel first order reaction model
152
168
The reaction constants and other parameters are shown in Table 6.4. The values kR for
all pHs ranged from 0.183 to 1.431 h-1 for the lowland water and from 0.135 to 1.294 h-1
for the upland water. In both waters, the rate of chlorine decay during the first phase of
the reaction was the fastest at pH 7, but no explanation was found as to why the rate of
chlorine decay was the fastest at pH 7. The values kS ranged between 0.001 and 0.004
and were not affected by the pH. The constants of rapid first order decay rate (kR) were
always greater than the constants of slower first order decay rate (kS). This is in line
with Gang et al. (2003), who also reported kR being 90 to 150 times larger than kS.
Here, for example, at pH 7, kR was about 350 and 430 times larger than kS for lowland
and upland water respectively. Then the ratios kR/kS were the greatest at pH 6 and then
pH 8 for both waters, indicating that the pH had an impact on the chlorine decay rate.
The f values, which represent the fraction of chlorine attributed to the rapid reaction,
ranged from 0.307 to 0.131, with the greatest values at pH 6, then pH 7 and pH 8.
Therefore, the chlorine consumption during the rapid reaction increased with decreasing
pH. Between 15 to 30% of the chlorine consumed was attributed to the rapid reaction,
which suggested that between 70 to 85% of the NOM reactivity was attributed to the
slow reaction, which agrees with Gang et al. (2003), who also reported a value of 70%
for three different fractions.
Table 6.4 Lowland and upland water chlorine decay constants for parallel first order
reaction model
Waters and conditions
LWa – pH = 6 – T = 20°C
LW – pH = 7 – T = 20°C
LW – pH = 8 – T = 20°C
UWb – pH = 6 – T = 20°C
UW – pH = 7 – T = 20°C
UW – pH = 8 – T = 20°C
a
kR (h-1)
0.874
1.431
0.183
0.135
1.294
0.351
kS (h-1)
0.004
0.004
0.003
0.001
0.003
0.003
f
0.209
0.185
0.159
0.307
0.182
0.131
R2
0.996
0.997
0.971
0.956
0.995
0.975
Lowland water; b Upland water.
As a summary, it was found here that:
 Data fitted well the parallel first order reaction model and this at three pHs.
 The first order rate constant for the rapid reaction kR was the largest at pH 7,
meaning that the chlorine decay was the fastest at pH 7.
153
 The pH did not impact on the second part of the reaction, which is the slowest
one.
 The chlorine consumption during the rapid reaction increased with decreasing
pH.
6.3.4.2 Chlorine decay to predict THM4 and HAA5
The coefficients obtained from the chlorine decay model were used previously by Gang
et al. (2002) to predict THM4 and HAA9. Here, it was decided to find out if the
coefficients found could be used to predict THM4 and HAA5. Therefore, on the basis on
the chlorine decay model, DBP formation can be expressed as follow:




THM 4  αC 0 1 - fe  k R t  (1  f)e  k S t ,
HAA 5  βC 0 1 - fe  k R t  (1  f)e  k St ,
Equation 6.3
Equation 6.4
where α = the THM4 yield coefficient, defined as the ratio of the concentration of THM4
(µg/L) to the concentration of chlorine consumed (mg/L), and β = the HAA5 yield
coefficient, defined as the ratio of the concentration of HAA5 (µg/L) to the
concentration of chlorine consumed (mg/L). Gang et al. (2002) presented the model
mathematical development from the chlorine decay to the model for DBPs. Here, the
differences were the number of HAAs to be modelled and the experimental conditions.
Gang et al. (2002) modelled nine HAAs. Here an attempt was made to determine if the
model could be applied for the five HAAs measured and at pH 6, 7 and 8.
Previously, in Section 6.3.2.4, it has been shown that the formation of THMs and HAAs
was characterised by an initial rapid reaction during the first few hours, followed by a
steady rate of increase over the 168 hours of contact time with chlorine. This was
observed here at any pHs (Figure 6.15 A + B and Figure 6.16 A + B). The impact of pH
on THMs and HAAs was explained in Section 6.3.2.1. The aim of this section here was
to fit the THM and HAA data to their respective model formation (Equation 6.3 and
6.4). The modelled data represented with lines in Figure 6.15 A + B and Figure 6.16 A
+ B showed that in general the models followed the formation of HAAs and THMs. The
coefficients of correlation R2, presented in the Table 6.5, were found between 0,755 and
154
0.969, with only two R2 being below 0.900. Gang et al. (2002) found that the formation
of THMs and HAAs was accurately simulated by the chlorine demand model in treated
waters. Here, the model could fit approximately HAAs and THMs for the two waters
and the three pHs studied.
90
LW - pH 6
LW - pH 7
180
A
LW - pH 8
70
UW - pH 6
160
Models
80
60
50
40
30
20
10
0
UW - pH 7
B
UW - pH 8
Models
140
120
100
80
60
40
20
0
0
24
48
72
96
120
144
168
0
24
48
72
96
120
144
168
Figure 6.15 HAA5 concentration and formation model predictions for (A) lowland water
(LW) and (B) upland water (UW), both at pH 6, 7 and 8
160
LW - pH 6
100
A
LW - pH 8
Models
120
100
80
60
40
20
UW - pH 6
90
140
LW - pH 7
UW - pH 7
B
UW - pH 8
Models
80
70
60
50
40
30
20
10
0
0
0
24
48
72
96
120
144
168
0
24
48
72
96
120
144
168
Figure 6.16 THM4 concentration and formation model predictions for (A) lowland
water (LW) and (B) upland water (UW), both at pH 6, 7 and 8
The THM4 and HAA5 coefficient yield (α and β respectively), both on a mass basis are
shown in Table 6.5. The THM4 coefficient yield α for all waters at three pHs ranged
between 16.24 to 36.84 µg THM4 per mg chlorine consumed. This is in line with Gang
et al. (2002), who reported the average THM4 yield coefficient of the treated waters to
be 30 µg THM per mg chlorine consumed. It was observed that increasing the pH
increased α in the lowland water, whereas the pH showed limited impact on α from the
upland water, which reinforced this statement already found in Section 6.3.2.1.
155
The HAA5 coefficient yield β found were between 19.01 and 51.80 µg HAA5 per mg
chlorine consumed. The value 19.01 µg HAA5 per mg chlorine consumed obtained here
at pH 8 is in line with the value of 18 µg HAA9 per mg chlorine consumed reported by
Gang et al. (2002), in their treated water at pH 8. While it was concluded previously that
the pH impact was equivocal on HAA yield, it was interesting to note here that β
decreased with increasing pH, as it was reported by previous research that HAA
formation decreased with increasing pH (Liang and Singer, 2003). Moreover, the
formation of HAA5 per mg of chlorine consumed was predominant in the upland water,
indicating more active precursors reacting with chlorine than in the lowland water,
which explained the greatest concentration of HAAs in the upland water.
Table 6.5 THM4 and HAA5 yield coefficients at three pHs and model coefficient of
correlation R2
a
α
(R2)
pH 6
16.24
(0.932)
Lowland water
pH 7
30.54
(0.953)
pH 8
36.84
(0.755)
pH 6
23.36
(0.959)
Upland water
pH 7
29.75
(0.934)
pH 8
24.41
(0.912)
βb
(R2)
26.80
(0.969)
26.50
(0.950)
19.01
(0.954)
51.80
(0.911)
45.19
(0.953)
42.90
(0.865)
a
α corresponds to µg THM4/mg Cl2 consumed; b β corresponds to µg HAA5/mg Cl2 consumed.
As a summary, it was found here that:
 THM4 and HAA5 were simulated by the chlorine decay model. In general, the
models followed approximately the DBP formation trend.
 The pH had a greater impact on α from the lowland water than on α from the
upland water.
 The coefficient yield β, which corresponds to µg HAA5 per mg chlorine
consumed, decreased with increasing pH.
6.4
CHAPTER SUMMARY
In Chapter 5, it was shown that HAAs and THMs were present in all the waters
surveyed and therefore, this extensive study on HAAs and THMs was designed to
determine the sensitivity of DBP formation to differences in water character and the
impact of water treatment parameters. The main results observed here were:
156
 The greatest pH impact was observed in the formation of THMs in the lowland
water. Although THM formation was significantly affected in both waters,
HAA5 formation did not exhibit a strong pH effect. When changing the pH, the
lowland water behaved as predicted in the reported literature with regard to
HAA formation. However the behaviour of the upland water did not follow the
same pattern. The differences in the precursor characteristics may account for
these observations.
 Addition of bromide to the water leads to a higher percentage of brominated
HAAs and THMs. Total THM4 increased, while total HAA5 decreased. The
impact on HAAs will vary depending on the number of brominated species
measured.
 A reduction in temperature resulted in a major decrease in DBP formation.
 The use of preformed chloramines reduced considerably the formation of HAAs
and THMs in UK treated waters.
 Fractions data varied considerably between water sources. The hydrophilic base
fraction from the lowland water was the most active towards the formation of
HAAs, whereas this was the hydrophobic acid for the upland water. The fraction
data obtained previously are based on the level of HAAs formed per mg of
carbon. However, each fraction contributed to a different percentage of the mass
distribution of NPOC and, therefore, it was necessary to consider the real
percentage of the HAA9. Therefore after reconstitution to their real contribution
in bulk water, it was concluded that the primary fractions to target for removal
were the hydrophilic acid + neutral and the hydrophobic acid for the lowland
and the upland water respectively.
 The chlorine decay data could fit a parallel first order reaction model that could
be used to predict the formation of HAAs and THMs.
 With the chlorine decay, the first order rate kR was the fastest at pH 7 for the first
part of the reaction, but the pH did not impact the slower part of the reaction.
Incorporation of chlorine during the first part of the reaction increased with
decreasing pH.
157
 The impact of pH was greater on α found in the lowland water than α from the
upland water. The coefficient yield β, which corresponds to µg HAA5 per mg
chlorine consumed, decreased with increasing pH.
As a general conclusion, significant water quality and treatment factors influence the
relative distribution and speciation of HAAs and THMs. Although some of the trends
followed the literature, some contradict previous findings. This study showed that water
utilities should monitor HAAs and THMs throughout their treatment plant, because the
formation of HAAs and THMs is highly specific to the water source.
158
Chapter 7
RECOMMENDATIONS FOR WATER UTILITIES
The present study has extended the knowledge of HAAs and other DBPs in treated
waters from around the UK. The findings of this study show that levels of HAAs and
other DBPs measured in treated waters vary considerably as a function of the
geographical location of the water treatment works. The work leads to a number of
recommendations based around key questions raised by water companies throughout the
duration of the project:
1. Question: What is the most appropriate procedure for measuring HAAs?
Answer: The most appropriate method is based on the two stage process of
methylation of the HAAs with acidic methanol followed by quantification using
GC/ECD. The identified method provides the only approach with a suitable
detection limit (1.435 g/L) in relation to legislative levels. The most common
alternative method, which utilises an IC, has a detection level that is not
sufficiently sensitive to be useful in practical situations; LOD could not be
determined with GC/MS. The GC/ECD method requires around 6 hours of
preparation time plus 30 minutes of GC/ECD analysis per sample. For analysis
of HAAs with this method consideration must be given to the validation of the
method with the specific machine being used in order to determine LOD and
peak retention time as this can vary between analytical devices.
2. Question: What is the most appropriate procedure for measuring semi-volatile
DBPs (if required in future)?
Answer: The most appropriate method is based on a two stage process of LLE
followed by quantification using GC/ECD as it is the most rapid and cost
effective of the available methods. The identified method has been demonstrated
to be suitable for i-THMs, HANs, HKs, HAs and HNMs. The method requires
around 2 hours of preparation time plus 30 minutes of GC/ECD analysis per
sample. Semi-volatile DBPs are easily degradable, therefore, the quenching
159
Chapter 7 – Recommendations for water utilities
agent should be carefully chosen amongst those known to be inert towards DBPs
such as ammonium chloride and ascorbic acid. As for HAAs, individual method
validation is required for semi-volatile DBPs when using a new instrument as
LOD and peak retention time can vary between analytical devices. The
identified method can support more compounds such as other i-DBPs, other
HANs, other HKs etc. but would require the use of GC/MS for confirmation if
use of a different GC column and/or incorporation of new compounds unknown
in the literature.
3. Question: What should be included in future DBP surveys?
Answer: Future surveys should focus on THM4 and HAA9 to reflect future
regulation. In addition to measuring these DBPs, samples should be analysed for
NPOC, UV254, and SUVA254 and bromide concentrations. The bromide
concentration can be used to establish likely speciation based on the threshold
limits identified in this work of 75 µg/L for THMs and 100 µg/L for HAAs. In
the future when semi-volatile DBPs are included in regulations, further analysis
will be required including the iodide levels to compare to threshold limits. This
will be equivalent to the bromide case reported above.
Attempts to find correlations between parameters should be restricted to
consistent water sources (e.g. lowland or groundwater) or by process, such as
coagulation and pre-ozonation to reduce the uncertainty which is inherent in
such work due to the complex nature of the organic make-up.
4. Question: Does the fact that a water meets the THM regulatory level mean that it
will be likely to meet the future regulations for HAAs?
Answer: NO. A key finding of the current work is a disparity in the correlation
between THMs and HAAs such that compliance to the regulation for one group
of DBPs does not indicate that the water will also be compliant to the other
group of DBPs. As such, waters that have not previously had a problem with
THM compliance may be currently unable to meet HAA compliance levels
without changing the operating strategy or the need for additional treatment
processes.
160
5. Question: How can the formation of DBPs be controlled?
Answer: The work presented has identified a number of parameters in relation to
existing treatment works that can be considered when attempting to manage
DBP formation:
(a) The switch to preformed monochloramine provides a consistent reduction in
DBPs formation within all classes of compounds studied: HAAs, THMs and
emerging DBPs, one exception being 1.1-DCP.
(b) The kinetics of the chlorine reactions indicate that reducing the contact time
of free chlorine will consistently reduce DBP formation levels.
(c) Alteration of the pH has a significant impact on DBP levels. THMs are
consistently reduced in terms of total THMs by operating under more acidic
conditions. The impact is less significant and becomes source dependent
with respect to HAAs. Based on one sample, a reduction of approximately
15% in total HAAs can be expected in the lowland water when the pH is
changed from 6 to 8, whereas this is an increase of approximately 15% in
total HAAs in the upland water. As such, pH control of HAAs by switching
to alkaline conditions is only suitable for lowland water.
(d) Temperature has a consistent impact with reduced DBPs under lower
temperature conditions. Whilst temperature is not a control variable in
practice it should be recognised that, with all other factors considered,
maximum DBPs will form in the summer. In practice this is complicated as
load and character of the organics change with season complicating the
impact of this parameter.
(e) Removal of specific precursors will reduce DBPs significantly. Whilst
individual precursor molecules are unknown, the work has identified key
organic fractions to target for different water sources. The specific fractions
are the hydrophilic acid + neutral fraction for lowland water and the
hydrophobic acid fraction for upland water. Focussing treatment
optimisation and improvements on these fractions should maximise the
161
impact of any changes in terms of reduction of DBPs and should guide
future technology selection.
162
Chapter 8
CONCLUSIONS
This project has extended the knowledge of DBP-FP from UK treated waters.
Specifically the work detailed in the thesis has identified the most suitable method for
the quantification of HAAs, and also developed an analytical method for the
determination of semi-volatile DBPs. Application of these methods to a survey of 11
different waters from around the UK indentified the complex interaction between the
variation in water quality characteristics, treatment flow sheet and the quantity and
speciation of the DBPs formed. Results presented in the thesis have voluntary been
quantified in µg/L in order to compare directly with the regulations set-up and to come.
However, it must be considered that adjusting these data to molar concentrations may
have an impact on the trends reported here.
Analysis of the resulted presented in this thesis leads to the following overall
conclusions:
1. The monitoring of HAAs and other DBPs can be achieved with reliable and
reproducible methods. The most suitable analytical method for the measurement
of nine HAAs in UK treated water is the methylation of the HAAs followed by
quantification with GC/ECD, with a detection limits for the sum of the nine
HAAs being 1.435 µg/L. Alternative methods such as the IC are not suitable as
their detection limits are too high; LODs for IC are 2.6 to 31.9 µg/L for
individual HAA. The most rapid and cost-effective method for the monitoring of
16 semi-volatile DBPs is LLE followed by quantification using GC/ECD, with a
detection limit for the sum of the 16 compounds being 0.907 µg/L. As a general
overview, the best device for the analysis of DBPs in treated waters is the
GC/ECD. In general, the use of FP test was shown to be a very useful mean of
understanding the formation of DBPs. However, the use of FP test was found
unsuitable for the quantification of i-THMs as the chlorine, used in excess, is
likely to limit their formation.
163
Chapter 8 – Conclusions
2. HAAs and semi-volatile DBPs including THMs, i-THMs, HANs, HKs, HAs and
HNMs were detected in UK treated waters. The 11 surveyed waters have the
potential to form a maximum of 68 µg/L for HAA9, 66 µg/L for THM4, 0.73
µg/L for i-THMs, 5.5 µg/L for HANs, 3.9 µg/L for HKs, 9.5 µg/L for HAs and
3.4 µg/L for HNMs. Improving the control of HAAs and other DBPs can be
achieved by using monochloramine instead of free chlorine, as it has been found
that, in general, a decrease in DBP concentrations has been observed when
shifting from chlorine to monochloramine, the one exception being 1,1-DCP.
The prevalent classes of DBPs were HAAs and THMs; HAs were found to be
the third major group. In general, the concentrations of THMs correlated well
with HAAs, and in particular the levels of TCM were similar to the levels of
TCAA supporting the hypothesis that they share similar precursor material.
Analysis of the correlation revealed that, in general, the regulatory limit of 100
µg/L for the THM4 would fail a regulation of 80 µg/L for the nine HAAs.
3. Modelling of the formation of HAAs and THMs using a parallel first order
reaction model, fitted to the chlorine decay data provided a good method of
prediction of the DBP formation levels. Thus, the rate of DBP formation is
characterised by a general trend of an initial rapid reaction during the first few
hours, followed by a steady rate of increase over the 168 hours of contact time
with chlorine.
4. The work outlined in the thesis identified a number of parameters which have
particular relevance when considering the formation of HAAs and THMs in
treated waters and it can be concluded that these factors should always be
considered when investigating or measuring DBPs:

The data identified threshold bromide level beyond which the
speciation of HAAs shifts toward predominantly brominated species.
The threshold levels were identified as 100 µg/L for HAAs and 75 µg/L
For THMs.

Management of the balance of THM and HAA is, in part, possible
through pH control as THM formation is significantly affected by the
pH whereas, HAA5 formation is less strongly linked. The impact in
164
Chapter 8 – Conclusions
relation to HAAs is further complicated by a switch in the impact of pH
by water sources. This is illustrated by a decrease of 15% in HAAs for a
lowland water and an increase of 14% as the pH switches from slightly
acidic to slightly alkaline.

Another important parameter affecting HAAs is the precursor make-up.
The use of bulk water characteristics, such as NPOC or SUVA, as
surrogates does not provide suitable means of controlling DBPs. This is
illustrated in its simplest form by the fact that two waters with similar
water characteristics can form different level of DBPs. Instead
characterisation by means of non specific ion exchange fractionation
indentifies the key group of compounds to target for each source water
type. Specifically, from the data presented it can be concluded that the
primary fractions to target for removal in order to minimise DBP
formation are the hydrophilic acid + neutral for lowland water and the
hydrophobic acid fraction for upland water.

Temperature has a consistent impact with a decreasing DBP formation
in lower temperatures and thus seasonal consideration will always be
important when considering strategies to meet legislative compliance
levels but must be viewed in combination with the other factors above.
165
166
Chapter 9
FURTHER WORK
The results of the present work highlight the complexity of the formation of HAAs and
other DBPs in UK treated waters and show that we are still far from a full
understanding. Furthermore, all results were obtained from controlled laboratory scaled
experiments. Further research should apply the laboratory findings to drinking waters
from the distribution systems, and focus on:
 Monitoring semi-volatile DBPs: The results found here showed that all the
waters studied had not only HAAs and THMs, but also i-THMs, HANs, HKs,
HAs and HNMs. This is the first time these compounds have been detected in
UK treated waters and, therefore, more work should be done to investigate the
real potential of drinking water to form these DBPs. The work should use
samples from different points of a distribution system to determine levels and
stability of these compounds. Seasonal variation could also be investigated in
order to determine the impact of NOM variation on the level of DBPs.
 Expand the number of DBPs monitored: Whereas 25 compounds have been
measured in the present study, about 100 DBPs are currently studied in the US
for their moderate occurrence levels and their potential toxicity. Therefore,
further work should focus on targeting new compounds, such as the iodo-acids,
the
halofuranones
(MX
[3-chloro-4-(dichloromethyl)-5-hydroxy-2-(5H)-
furanone] and brominated MX DBPs), N-nitrosodimethylamine (NDMA) and
other nitrosamines, etc. and set-up analytical methods for their measurement.
 Set-up and monitor dissolved organic nitrogen (DON) levels, removal and
reactivity with chlorine: There is evidence that the DON present in NOM acts as
precursors to nitrogen-containing DBPs. This is important, because N-DBPs
have been reported to be more toxic than chlorinated or brominated DBPs.
167
Chapter 9 – Further work
 Set-up and monitor total organic halogen (TOX), with its halogen-specific
fraction: total organic chlorine (TOCl), total organic bromine (TOBr) and total
organic iodine (TOI): TOX screening can be used to quantify the chlorinated,
brominated and iodinated organic compounds in drinking waters. By
determining TOX and the known DBP concentration, it would then be possible
to estimate the percentage of unknown DBPs in UK drinking waters.
 Understanding the pathways and links between DBPs: Further research should
focus on progressive degradation of one species into a new species. For
example, this work could be carried out by spiking a known amount of TCM
into pure water and by adding excess chlorine for a determined contact time. At
the end of the incubation period, the sample could be analysed by GC/ECD for
analysis of DBPs. In this way, the DBP degradation or transformation could be
easily determined, as well as the link between species.
 CT values when using chloramines as a secondary disinfectant: A long prechlorination time results in higher concentration of DBPs. Therefore, by
optimising the CT values, a possible decrease of DBPs could be expected.
168
Chapter 10
REFERENCES
Adin, A., Katzhendler, J., Alkaslassy, D., and Rav-Acha, C. (1991). Trihalomethane
formation in chlorinated drinking water: A kinetic model. Water Research, 25(7), p.
797-805.
Allgeier, S. C., and Summers, R. S. (1995). Evaluating NF for DBP control with the
RBSMT. Journal of the American Water Works Association, 87(3), p. 87-99.
American Public Health Association / American Water Works Association / Water
Environment Federation (1998). Standard Methods for the Examination of Water and
Wastewater, 20th ed. APHA/AWWA/WEF, Washington, DC, US.
American Public Health Association / American Water Works Association / Water
Environment Federation (1992). Standard Methods for the Examination of Water and
Wastewater, 18th ed. APHA/AWWA/WEF, Washington, DC, US.
Amy, G. L., Siddiqui, M., Ozekin, K., Zhu, H.-W., and Wang, C. (1998). Empirical
based models for predicting chlorination and ozonation by-products: Haloacetic acids,
chloral hydrate, and bromate, Environmental Protection Agency report no. CX 819579,
US.
Amy, G. L., Siddiqui, M., Zhai, W., and Debroux, J. (1994). Survey on bromide in
drinking water and impacts on DBP formation, American Water Works Association
Research Foundation report no. 90662, Denver, CO, US.
Amy, G. L., Tan, L., and Davis M. K. (1991). The effect of ozonation and activated
carbon adsorption on trihalomethane speciation. Water Research, 25(2), p. 191-202.
Amy, G.L., Chadik, P.A., and Chowdhury, Z.K. (1987). Developing models for
predicting THM formation potential and kinetics. Journal of the American Water Works
Association, 79(7), p. 89-97.
Ates, N., Kaplan, S. S., Sahinkaya, E., Kitis, M., Dilek, F. B., and Yetis, U. (2007).
Occurrence of disinfection by-products in low DOC surface waters in Turkey. Journal
of Hazardous Materials, 142(1-2), p. 526-534.
Bellar, T. A., Lichtenberg, J. J., and Kroner, R. C. (1974). Occurrence of organohalides
in chlorinated drinking waters. Journal of the American Water Works Association,
66(12), p. 703-706.
169
Chapter 10 – References
Bichsel, Y., and Von Gunten, U. (2000). Formation of iodo-trihalomethanes during
disinfection and oxidation of iodide-containing waters. Environmental Science and
Technology, 34(13), p. 2784-2791.
Bichsel, Y., and Von Gunten, U. (2000a). Hypoiodous acid: Kinetics of the buffercatalyzed disproportionation. Water Research, 34(12), p. 3197-3203.
Bichsel, Y., and Von Gunten, U. (1999). Oxidation of iodide and hypoiodous acid in the
disinfection of natural waters. Environmental Science and Technology, 33(22), p. 40404045.
Boorman, G. A., Dellarco, V., Dunnick, J. K., Chapin, R. E., Hunter, S., Hauchman, F.,
Gardner, H., Mike, C., and Sills, R. C. (1999). Drinking water disinfection byproducts:
Review and approach to toxicity evaluation. Environmental Health Perspectives,
107(1), p. 207-217.
Bove, F., Shim, Y., and Zeitz, P. (2002). Drinking water contaminants and adverse
pregnancy outcomes: A review. Environmental Health Perspectives, 110(1), p. 61-74.
Bull, R. J., Sanchez, I. M., Nelson, M. A., Larson, J. L., and Lansing, A. J. (1990). Liver
tumor induction in B6C3F1 mice by dichloroacetate and trichloroacetate. Toxicology,
63(3), p. 341-359.
Bull, R. J., and Robinson, M. (1986). Carcinogenic activity of haloacetonitrile and
haloacetone derivatives in the mouse skin and lung. In: Water Chlorination: Chemistry,
Environmental Impact and Health Effects, edited by Jolley, R.L. Chelsea, MI, US, vol.
5, p. 221-227.
Cancho, B., Ventura, F., Galceran, M., Diaz, A., and Ricart, S. (2000). Determination,
synthesis and survey of iodinated trihalomethanes in water treatment processes. Water
Research, 34(13), p. 3380-3390.
Cantor, K. P., Lynch, C. F., Hildesheim, M. E., Dosemeci, M., Lubin, J., Alavanja, M.,
and Craun, G. (1998). Drinking water source and chlorination byproducts I. Risk of
bladder cancer. Epidemiology, 9(1), p. 21-28.
Carlson, M., and Hardy, D. (1998). Controlling DBPs with monocholoramine. Journal
of the American Water Works Association, 90(2), p. 95-106.
Carrero, H., and Rusling, J. F. (1999). Analysis of haloacetic acid mixtures by HPLC
using an electrochemical detector coated with a surfactant-nafion film. Talanta, 48(3),
p. 711-718.
Chang, E. E., Lin, Y. P., and Chiang, P. C. (2001). Effects of bromide on the formation
of THMs and HAAs. Chemosphere, 43(8), p. 1029-1034.
Chang, E.E, Chao, S., Chiang, P.C., and Lee, J. (1996). Effects of chlorination on THM
formation in raw water. Toxicological & Environmental Chemistry, 56, p. 211-225.
170
Chinn, R., Lee, T., Krasner, S. W., Dale, M., Richardson, S. D., Pressman, J., Speth, T.,
Miltner, R. J. and Simmons, J. E. (2007). Solid-phase extraction of 35 DBPs with
analysis with GC/ECD and GC/MS. In: Proceedings of the American Water Works
Association Water Quality Technology Conference, November 4-8, Charlotte, NC, US.
Chowdhury, Z. K., and Amy, G. L. (1999). Modelling disinfection by-product
formation. In: Formation and control of disinfection by-products in drinking water,
edited by P.C. Singer, American Water Works Association, Denver, CO, US, p. 53-64.
Christman, R. F., Norwood, D. L., Millington, D. S., and Johnson, J. D. (1983). Identity
and yields of major halogenated products of aquatic fulvic acid chlorination.
Environmental Science and Technology, 17(10), p. 625-628.
Clark, R. M., Thurnau, R. C., Sivaganesan, M., and Ringhand, P. (2001). Predicting the
formation of chlorinated and brominated by-products. Journal of Environmental
Engineering, 127(6), p. 493-501.
Clark, R. M. (1998). Chlorine demand and TTHM formation kinetics: A second order
model. Journal of Environmental Engineering, 124(1), p. 16-24.
Coleman, W. E., Munch, J. W., Kaylor, W. H., Streicher, R. P., Ringhand, H. P., and
Meier, J. R. (1984). Gas chromatography/mass spectroscopy analysis of mutagenic
extracts of aqueous chlorinated humic acid. A comparison of the byproducts to drinking
water contaminants. Environmental Science and Technology, 18(9), p. 674-681.
Collins, M. R., Amy, G. L., and Steelink, C. (1986). Molecular weight distribution,
carboxylic acidity, and humic substances content of organic matter: Implications for
removal during water treatment. Environmental Science and Technology, 20(10), p.
1028-1032.
Cortvriend, J. (2008). Establishment of a list of chemical parameters for the revision of
the Drinking Water Directive, ENV.D.2/ETU/2007/0077r.
Cowman, G. A., and Singer, P. C. (1996). Effect of bromide ion on haloacetic acid
speciation resulting from chlorination and chloramination of aquatic humic substances.
Croué, J.-P., Korshin, G. V., and Benjamin, M. M. (2000). Characterization of natural
organic matter in drinking water, American Water Works Association Research
Foundation report no. 90780, Denver, CO, US.
Croué, J.-P., Martin, B., Deguin, A., and Legube, B. (1993). Isolation and
characterisation of dissolved hydrophobic and hydrophilic organic substances of a
reservoir water. In: Natural Organic Matter in Drinking Water, Origin,
Characterization and Removal, Proceedings of the American Water Works Association
Research Foundation NOM Workshop, September 9-22, Chamonix, France, p. 73-81.
Croué, J.-P., and Reckhow, D. A. (1989). Destruction of chlorination byproducts with
sulfite. Environmental Science and Technology, 23(11), p. 1412-1419.
171
Daniel, F. B., Schenck, K. M., and Mattox, J. K. (1986). Genotoxic properties of
haloacetonitriles: Drinking water by-products of chlorine disinfection. Fundamental and
Applied Toxicology, 6(3), p. 447-453.
Deborde, M., and Von Gunten, U. (2008). Reactions of chlorine with inorganic and
organic compounds during water treatment - Kinetics and mechanisms: A critical
review. Water Research, 42(1-2), p. 13-51.
Diehl, A. C., Speitel, G. E., Jr., Symons, J. M., Krasner, S. W., Hwang, C. J., and
Barrett, S. E. (2000). DBP formation during chloramination. Journal of the American
Water Works Association, 92(6), p. 76-90.
Dojlido, J., Zbiec, E., and Swietlik, R. (1999). Formation of the haloacetic acids during
ozonation and chlorination of water in Warsaw waterworks (Poland). Water research,
33(14), p. 3111-3118.
Doyle, T. J., Zheng, W., Cerhan, J. R., Hong, C., Sellers, T. A., Kushi, L. H., and
Folsam, A. R. (1997). The association of drinking water source and chlorination byproducts with cancer incidence among postmenopausal women in Iowa: A prospective
cohort study. American Journal of Public Health, 87(7), p. 1168-1176.
Drinking
Water
Inspectorate
(DWI)
http://www.dwi.gov.uk/31/pdf/soslist06.pdf.
(2009),
available
at:
Drinking
Water
Inspectorate
(DWI)
www.dwi.gov.uk/regs/pdf/GuidanceMay05.pdf.
(2005),
available
at:
Duirk, S. E., and Valentine, R. L. (2006). Modeling dichloroacetic acid formation from
the reaction of monochloramine with natural organic matter. Water Research, 40(14), p.
2667-2674.
Duirk, S. E., Gombert, B., Croué, J.-P., and Valentine, R. L. (2005). Modeling
monochloramine loss in the presence of natural organic matter. Water Research, 39(14),
p. 3418-3431.
Dunnick, J. K., Haseman, J. K., Lilja, H. S., and Wyand, S. (1985). Toxicity and
carcinogenicity of chlorodibromomethane in Fischer 344/N rats and B6C3F1 mice.
Fundamental and Applied Toxicology, 5(6 I), p. 1128-1136.
Edzwald, J. K., and Tobiason, J. E. (1999). Enhanced coagulation: US requirements and
a broader view. Water Science and Technology, 40(9), p. 63-70.
El-Dib, M. A., and Ali, R. K. (1995). THMs formation during chlorination of raw Nile
river water. Water Research, 29(1), p. 375-378.
Fearing, D. A., Goslan, E. H., Banks, J., Wilson, D., Hillis, P., Campbell, A. T., and
Parsons, S. A. (2004). Staged coagulation for treatment of refractory organics. Journal
of Environmental Engineering, 130(9), p. 975-982.
172
Fischbacher, C. (2000). Quality assurance in analytical chemistry. In: Encyclopedia of
Analytical Chemistry: Applications, Theory and Instrumentation, edited by Meyers,
R.A. Tarzana, CA, US, vol. 15, p. 13563-13585.
Fleischacker, S. J., and Randtke, S. J. (1983). Formation of organic chlorine in public
water supplies. Journal of the American Water Works Association, 75(3), p. 132-138.
Fuge, R., and Johnson, C. C. (1986). The geochemistry of iodine - A review.
Environmental Geochemistry and Health, 8(2), p. 31-54.
Gallard, H., and Von Gunten, U. (2002). Chlorination of natural organic matter:
Kinetics of chlorination and THM formation. Water Research, 36(1), p. 65-74.
Gang, D.D., Clevenger, T. E, and Banerji, S. K. (2003). Relationship of chlorine decay
and THMs formation to NOM size. Journal of Hazardous Materials, 96(1), p. 1-12.
Gang, D.D., Segar, R. L., Jr., Clevenger, T. E., and Shankha, K. B. (2002). Using
chlorine demand to predict THM and HAA9 formation. Journal of the American Water
Works Association, 94(10), p. 76-86.
Garcia-Villanova, R. J., Garcia, C., Gomez, J. A., Garcia, M. P., Ardanuy, R. (1997).
Formation, evolution and modelling of trihalomethanes in the drinking water of a town:
II. In the water distribution system. Water Research, 31(6), p. 1405-1413.
Glezer, V., Harris, B., Tal, N., Iosefzon, B., and Lev, O. (1999). Hydrolysis of
haloacetonitriles: Linear free energy relationship. Kinetics and products. Water
Research, 33(8), p. 1938-1948.
Golfinopoulos, S. K., and Nikolaou, A. D. (2005). Survey of disinfection by-products in
drinking water in Athens, Greece. Desalination, 176(1-3), p. 13-24.
Golfinopoulos, S. K., Arhonditsis, G. B. (2002). Multiple regression models: A
methodology for evaluating trihalomethane concentrations in drinking water from raw
water characteristics. Chemosphere, 47(9), p. 1007-1018.
Gonzalez, A. C., Krasner, S. W., Weinberg, H., and Richardson, S. D. (2000).
Determination of newly identified disinfection by-products in drinking water. In:
Proceedings of the American Water Works Association Water Quality Technology
Conference, November 5-9, Salt Lake City, UT, US.
Goslan, E. H., Krasner, S. W., Bower, M., Rocks, S. A., Holmes, P., Levy, L. S., and
Parsons, S. A. (2009). A comparison of disinfection by-products found in chlorinated
and chloraminated drinking waters in Scotland. Water Research (Submitted).
Goslan, E. H. (2003). Natural organic matter character and reactivity: Assessing
seasonal variation in a moorland water. Eng. D. Thesis, Cranfield University, UK.
173
Goslan, E. H., Fearing, D. A., Banks, J., Wilson, D., Hills, P., Campbell, A. T., and
Parsons, S. A. (2002). Seasonal variations in the disinfection by-product precursor
profile of a reservoir water. Journal of Water Supply: Research and Technology –
AQUA,, 51(8), p. 475-482.
Greiner, A. D., Obolensky, A., and Singer, P. C. (1992). Technical note: Comparing
predicted and observed concentrations of DBPs. Journal of the American Water Works
Association, 84(11), p. 99-102.
Guay, C., Rodriguez, M., and Sérodes, J. (2005). Using ozonation and chloramination to
reduce the formation of trihalomethanes and haloacetic acids in drinking water.
Desalination, 176(1-3), p. 229-240.
Gurol, M. D., Wowk, A., Myers, S., and Suffet, I. H. (1983). Kinetics and mechanism
of haloform formation: chloroform formation from trichloroacetone. In: Water
Chlorination: Environmental Impact and Health Effects, edited by Jolley, R.L. Ann
Arbor, MI, US, vol. 4, p. 269-284.
Heller-Grossman, L., Manka, J., Limoni-Relis, B., and Rebhun, M. (1993). Formation
and distribution of haloacetic acids, THM and TOX in chlorination of bromide-rich lake
water. Water research, 27(8), p. 1323-1331.
Hildesheim, M. E., Cantor, K. P., Lynch, C. F., Dosemeci, M., Lubin, J., Alavanja, M.,
and Craun, G. (1998). Drinking water source and chlorination byproducts II. Risk of
colon and rectal cancers. Epidemiology, 9(1), p. 29-35.
Hoigné, J., and Bader, H. (1988). The formation of trichloronitromethane (chloropicrin)
and chloroform in a combined ozonation/chlorination treatment of drinking water.
Water Research, 22(3), p. 313-319.
Hong, Y., Liu, S., Song, H., and Karanfil, T. (2007). HAA formation during
chloramination - Significance of monochloramine's direct reaction with DOM. Journal
of the American Water Works Association, 99(8), p. 57-59.
Hozalski, R. M., Zhang, L., and Arnold, W. A. (2001). Reduction of haloacetic acids by
Fe0: Implications for treatment and fate. Environmental Science and Technology,
35(11), p. 2258-2263.
Hua, G., and Reckhow, D. A. (2007). Comparison of disinfection byproduct formation
from chlorine and alternative disinfectants. Water Research, 41(8), p. 1667-1678.
Hua, G., Reckhow, D. A., and Kim, J. (2006). Effect of bromide and iodide ions on the
formation and speciation of disinfection byproducts during chlorination. Environmental
Science and Technology, 40(9), p. 3050-3056.
Huang, J., Graham, N., Templeton, M. R., Zhang, Y., Collins, C., and Nieuwenhuijsen,
M. (2008). Evaluation of Anabaena flos-aquae as a precursor for trihalomethane and
haloacetic acid formation. Water Science and Technology: Water Supply, 8(6), p. 653662.
174
Hwang, C. J., Amy, G. L., Bruchet, A., Croué, J.-P., Krasner, S. W., and Leenheer, J. A.
(2001). Polar NOM: Characterization, DBPs, treatment. American Water Works
Association Research Foundation report no. 90877, Denver, CO, US.
Hwang, C. J., Sclimenti, M. J., and Krasner, S. W. (2000). Disinfection by-product
formation reactivities of natural organic matter fractions of a low-humic water. In:
Natural Organic Matter and Disinfection By-products: Characterization and Control in
Drinking Water. American Chemical Society Symposium Series 761, edited by Sylvia
E. Barrett, Stuart W. Krasner and Gary L. Amy. Washington, DC, US, p. 173-187.
Jarvis, P., Parsons, S. A., and Smith, R. (2007). Modeling bromate formation during
ozonation. Ozone: Science and Engineering, 29(6), p. 429-442.
Jauhiainen, T., Moore, J., Perämäki, P., Derome, J., and Derome, K. (1999). Simple
procedure for ion chromatographic determination of anions and cations at trace levels in
ice core samples. Analytica Chimica Acta, 389(1-3), p. 21-29.
Kanokkantapong, V., Marhaba, T. F., Pavasant, P., and Panyapinyophol, B. (2006).
Characterization of haloacetic acid precursors in source water. Journal of
Environmental Management, 80(3), p. 214-221.
Karanfil, T., Hong, Y., and Song, H. (2008). HAA formation and speciation during
chloramination. In: Disinfection By-Products in drinking Water: Occurrence,
Formation, Health Effects, and Control. American Chemical Society Symposium Series
995, edited by Karanfil, T., Krasner, S. W., Westerhoff, P. and Xie, Y. Washington,
DC, US, p. 95–108.
Karanfil, T., Hong, Y., Song, H., and Orr, O. (2007). Exploring the pathways of HAA
formation during chloramination, American Water Works Association Research
Foundation report 91192, Denver, CO, US.
Kargalioglu, Y., McMillan, B.J., Minear, R.A., and Plewa, M.J. (2002). An analysis of
cytotoxicity and mutagenicity of drinking water disinfection by-products in Salmonella
typhimurium. Teratogen. Carcinogen. Mutagen., 22(2), p. 113–128.
Kim, H. C., and Yu, M. J. (2005). Characterization of natural organic matter in
conventional water treatment processes for selection of treatment processes focused on
DBPs control. Water Research, 39(19), p. 4779-4789.
Kim J., Chung Y., Shin D., Kim M., Lee Y., Lim Y., and Lee D. (2002). Chlorination
by-products in surface water treatment process. Desalination, 151(1), p. 1-9.
Kim, D. H., Choi, J. O., Kim, M., and Dai Woon, L. (2001). Determination of
haloacetic acids in tap water by capillary electrophoresis with direct UV detection.
Journal of Liquid Chromatography and Related Technologies, 24(1), p. 47-55.
King, W. D., Dodds, L., and Allen, A. C. (2000). Relation between stillbirth and
specific chlorination by-products in public water supplies. Environmental Health
Perspectives, 108(9), p. 883-886.
175
King, W. D., and Marrett, L. D. (1996). Case-control study of bladder cancer and
chlorination by-products in treated water (Ontario, Canada). Cancer Causes and
Control, 7(6), p. 596-604.
Kirmeyer, G., Martel, K., Thompson, G., Radder, L., Klement, W., LeChevallier, M.,
Baribeau, H., and Flores, A. (2004). Optimizing chloramine treatment. 2nd ed.
American Water Works Association Research Foundation, Denver, CO, US.
Koivusalo, M., Pukkala, E., Vartiainen, T., Jaakkola, J. J. K., and Hakulinen, T. (1997).
Drinking water chlorination and cancer - A historical cohort study in Finland. Cancer
Causes and Control, 8(2), p. 192-200.
Koudjonou, B., LeBel, G. L., and Dabeka, L. (2008). Formation of halogenated
acetaldehydes, and occurrence in Canadian drinking water. Chemosphere, 72(6), p. 875881.
Krasner, S. W., Sclimenti, M. J., Mitch, W., Westerhoff, P., and Dotson, A. (2007).
Using formation potential tests to elucidate the reactivity of DBP precursors with
chlorine versus with chloramines. In: Proceedings of the American Water Works
Association Water Quality Technology Conference, November 4-8, Charlotte, NC, US.
Krasner, S. W., Weinberg, H. S., Richardson, S. D., Pastor, S. J., Chinn, R., Sclimenti,
M. J., Onstad, G. D., and Thruston, A. D., Jr. (2006). Occurrence of a new generation of
disinfection byproducts. Environmental Science and Technology, 40(23), p. 7175-7185.
Krasner, S. W., Pastor, S., Chinn, R., Sclimenti, M. J., Weinberg, H. S., Richardson, S.
D., and Thruston, A. D., Jr. (2001). The occurrence of a new generation of DBPs
(beyond the ICR). In: Proceedings of the American Water Works Association Water
Quality Technology Conference, November 11-14, Nashville, TN, US.
Krasner, S. W. (1999). Chemistry of Disinfection By-Product Formation. In: Formation
and Control of Disinfection By-Products in Drinking Water, edited by Singer, P.C.,
American Water Works Association, Denver, CO, US, p. 27-52.
Krasner, S. W., McGuire, M. J., Jacangelo, J. G., Patania, N. L., Reagan, K. M., and
Marco Aieta, E. (1989). Occurrence of disinfection by-products in US drinking water.
Journal of the American Water Works Association, 81(8), p. 41-53.
La Guardia, M. J. (1996). EPA Method 552.2 ICR monitoring and the effects using
sodium bicarbonate. In: Proceedings of the American Water Works Association Water
Quality Technology Conference, November 17-21, Boston, MA, US.
Lee, W., Westerhoff, P., and Croué, J.-P. (2007). Dissolved organic nitrogen as a
precursor for chloroform, dichloroacetonitrile, N-nitrosodimethylamine and
trichloronitromethane. Environmental Science and Technology, 41(15), p. 5485-5490.
Leenheer, J. A., Noyes, T. I., Rostad, C. E., and Davisson, M. L. (2004).
Characterization and origin of polar dissolved organic matter from the Great Salt Lake.
Biogeochemistry, 69(1), p. 125-141.
176
Leenheer, J.A. (1981). Comprehensive approach to preparative isolation and
fractionation of dissolved organic carbon from natural waters and wastewaters.
Environmental Science and Technology, 15(5), 578-587.
Liang, L., and Singer, P. C. (2003). Factors influencing the formation and relative
distribution of haloacetic acids and trihalomethanes in drinking water. Environmental
Science and Technology, 37(13), p. 2920-2928.
Liu, Y., and Mou, S. (2003). Determination of trace levels of haloacetic acids and
perchlorate in drinking water by ion chromatography with direct injection. Journal of
Chromatography A, 997(1-2), p. 225-235.
Lou, J. C., and Chiang, P. C. (1994). A study of trihalomethanes formation in water
distribution system. Hazard Waste and Hazardous Materials, 11(2), p. 333-343.
Lovelock, J. E. (1974). The electron capture detector. Theory and practice. Journal of
Chromatography A, 99(C), p. 3-12.
Lovelock, J. E. (1958). A sensitive detector for gas chromatography. Journal of
Chromatography A, 1(C), p. 35-46.
Malcolm, R. L., and MacCarthy, P. (1992). Quantitative evaluation of XAD-8 and
XAD- 4 resins used in tandem for removing organic solutes from water. Environment
International, 18(66), p. 597-607.
Malliarou, E., Collins, C., Graham, N., and Nieuwenhuijsen, M. (2005). Haloacetic
acids in drinking water in the United Kingdom. Water Research, 39(12), p. 2722-2730.
Marhaba, T. F., and Van, D. (2000). The variation of mass and disinfection by-product
formation potential of dissolved organic matter fractions along a conventional surface
water treatment plant. Journal of Hazardous Materials, 74(3), p. 133-147.
McGuire, M. J., McLain, J. L., and Obolensky, A. (2002). Information Collection Rule
Data Analysis, American Water Works Association Research Foundation report,
Denver, CO, US.
McKnight, A., and Reckhow, D. A. (1992). Reactions of ozonation by-products with
chlorine and chloramines, In: Proceedings of the American Water Works Association
Annual Conference, Water Research, June 18-22 1992, Vancouver, BC, Canada, p. 399409.
McRae, B. M., LaPara, T. M., and Hozalski, R. M. (2004). Biodegradation of haloacetic
acids by bacterial enrichment cultures. Chemosphere, 55(6), p. 915-925.
Miller, J. W., and Uden, P. C. (1983). Characterization of nonvolatile aqueous
chlorination products of humic substances. Environmental Science and Technology,
17(3), p. 150-157.
177
Mitch, W. A., and Sedlak, D. L. (2002). Formation of N-nitrosodimethylamine
(NDMA) from dimethylamine during chlorination. Environmental Science and
Technology, 36(4), p. 588-595.
Nair, L. M., Saari-Nordhaus, R., and Anderson, J. M., Jr. (1994). Determination of
haloacetic acids by ion chromatography. Journal of Chromatography A, 671(1-2), p.
309-313.
Nikolaou, A. D., Lekkas, T. D., and Golfinopoulos, S. K. (2004). Kinetics of the
formation and decomposition of chlorination by-products in surface waters. Chemical
Engineering Journal, 100(1-3), p. 139-148.
Nissinen, T. K., Miettinen, I. T., Martikainen, P. J., and Vartiainen, T. (2002).
Disinfection by-products in Finnish drinking waters. Chemosphere, 48(1), p. 9-20.
Nokes, C. J., Fenton, E., and Randall, C.J. (1999). Modelling the Formation of
Brominated Trihalomethanes in Chlorinated Drinking Waters. Water Research, 33(17),
p. 3557-3568.
Oliver, B. G. (1983). Dihaloacetonitriles in drinking water: Algae and fulvic acid as
precursors. Environmental Science and Technology, 17(2), p. 80-83.
Oliver, B. G., and Thurman, E. M. (1983). Influence of aquatic humic substance
properties on trihalomethane potential. In: Water Chlorination: Environmental Impact
and Health Effects, edited by Jolley, R. L., Brungs, W. A., Cotruvo, J. A., Cumming, R.
B., Mattice, J. S., and Jacobs, V. A. Ann Arbor, MI, US, vol. 1, p. 231-241.
Parsons, S. A, Goslan, E. H., Rocks, S., Holmes, P., Levy, L. S., and Krasner, S. (2009).
Study into the formation of disinfection by-products of chloramination, potential health
implications and techniques for minimisation. Cranfield University, UK, available at
http://www.scotland.gov.uk/Resource/Doc/1057/0081130.pdf.
Parsons, S. A., and Jefferson B. (2006). Introduction to Potable Water Treatment
Processes, edited by Blackwell Publishing Ltd, Oxford, UK.
Parsons, S. A., Jefferson, B., Goslan, E. H., Jarvis, P. R., and Fearing, D. A. (2005).
Natural organic matter - The relationship between character and treatability. Journal of
Water Supply: Research and Technology – AQUA,, 4(5-6), p. 43–48.
Peters, R. J. B., Erkelens, C., De Leer, E. W. B., and De Galan, L. (1991). The analysis
of halogenated acetic acids in Dutch drinking water. Water research, 25(4), p. 473-477.
Peters, R. J. B., De Leer, E. W. B., and De Galan, L. (1990). Dihaloacetonitriles in
Dutch drinking waters. Water Research, 24(6), p. 797-800.
Plewa, M. J., Wagner, E. D., Muellner, M. G., Hsu, K.-M., and Richardson, S. D.
(2008). Comparative mammalian cell toxicity of N-DBPs and C-DBPs. In: Disinfection
By-Products in drinking Water: Occurrence, Formation, Health Effects, and Control.
American Chemical Society Symposium Series 995, edited by Karanfil, T., Krasner, S.
W., Westerhoff, P. and Xie, Y. Washington, DC, US, p. 36–50.
178
Plewa, M. J., Wagner, E. D., Jazwierska, P., Richardson, S. D., Chen, P. H., and
McKague, A. B. (2004). Halonitromethane drinking water disinfection byproducts:
chemical characterization and mammalian cell cytotoxicity and genotoxicity.
Plewa, M. J., Kargalioglu, Y., Vankerk, D., Minear, R. A., and Wagner, E. D. (2002).
Mammalian cell cytotoxicity and genotoxicity analysis of drinking water disinfection
by-products. Environmental and Molecular Mutagenesis, 40(2), p. 134-142.
Pope, P. G., Speitel, G. E., Jr., and Collins, M. R. (2006). Kinetics of dihaloacetic acid
formation during chloramination. Journal of the American Water Works Association,
98(11), p.107-120.
Pourmoghaddas, H., Stevens, A. A., Kinman, R. N., Dressman, R. C., Moore, L. A., and
Ireland, J. C. (1993). Effect of bromide ion on formation of HAAs during chlorination.
Journal of the American Water Resources Association, 85(1), p. 82-87.
Quevauviller, P. (2002). Quality Assurance for Water Analysis, edited by John Wiley
and Sons Ltd, New York, NY, US.
Rathbun, R. E. (1996). Regression equations for disinfection by-products for the
Mississippi, Ohio and Missouri Rivers. The Science of the Total Environment, 191(3), p.
235-244.
Ratnaweera, H., Hiller, N., and Bunse, U. (1999). Comparison of the coagulation
behaviour of different Norwegian aquatic NOM sources. Environmental International,
25(2-3), p. 347-355.
Reckhow, D. A., and Singer, P. C. (1990). Chlorination by-products in drinking Waters:
from formation potentials to finished water concentrations. Journal of the American
Water Works Association, 82(4), p. 173-180.
Reckhow, D. A., Singer, P. C., and Malcolm, R. L. (1990). Chlorination of humic
materials: Byproduct formation and chemical interpretations. Environmental Science
and Technology, 24(11), p. 1655-1664.
Reckhow, D. A., and Singer, P. C. (1985). Mechanisms of organic halide formation
during fulvic acid chlorination and implications with respect to preozonation. In: Water
Chlorination: Chemistry, Environmental Impact and Health Effects, edited by Jolley,
R.L. Chelsea, MI, US, vol. 5, p. 1229-1257.
Reckhow, D. A., and Singer, P. C. (1984). The removal of organic halide precursors by
preozonation and alum coagulation. Journal of the American Water Works Association,
76(4), p. 151-157.
Richardson, S. D., Plewa, M. J., Wagner, E. D., Schoeny, R., and DeMarini, D. M.
(2007). Occurrence, genotoxicity, and carcinogenicity of regulated and emerging
disinfection by-products in drinking water: A review and roadmap for research. Reviews
in Mutation Research, 636(1-3), p. 178-242.
179
Richardson, S. D. (2003). Disinfection by-products and other emerging contaminants in
drinking water. Trends in Analytical Chemistry, 22(10), p. 666-684.
Richardson, S. D., Simmons, J. E., and Rice, G. (2002). Disinfection byproducts: The
next generation. Environmental Science and Technology, 36(9), p. 198A-205A.
Richardson, S. D., Thruston, A. D., Jr., Caughran, T. V., Chen, P. H., Collette, T. W.,
Floyd, T. L., Schenck, K. M., Lykins Jr., B. W., Sun, G. R., and Majetich, G. (1999).
Identification of new drinking water disinfection byproducts formed in the presence of
bromide. Environmental Science and Technology, 33(19), p. 3378-3383.
Richardson, S. D. (1998). Drinking water disinfection by-products. In: The
Encyclopedia of Environmental Analysis and remediation, edited by Meyers, R.A.
Wiley, New York, NY, US, vol. 3, p. 1398-1421.
Rodriguez, M. J., Sérodes, J., and Levallois, P. (2004). Behavior of trihalomethanes and
haloacetic acids in a drinking water distribution system. Water.research, 38(20), p.
4367-4382.
Rodriguez, M. J, Sérodes, J.-B., and Morin M. (2000). Estimation of water utility
compliance with trihalomethane regulations using modelling approach. Journal of
Water Supply: Research and Technology – AQUA, 49(2), p. 57-73.
Rook, J. J. (1974). Formation of haloforms during chlorination of natural waters. Water
Treatment and Examination, 23(2), p. 234-243.
Sadiq, R., and Rodriguez, M. J. (2004). Disinfection by-products (DBPs) in drinking
water and predictive models for their occurrence: A review. Science of the Total
Environment, 321(1-3), p. 21-46.
Shorney, H. L., Radtke, S., Hargette, P., Knocke, W., Dietrich, A., Hoehn, R., and
Long, B. (1999). Removal of DBP precursor by enhanced coagulation and lime
softening, American Water Works Association Research Foundation report, Denver,
CO, US.
Siddiqui, M., Amy, G., Ryan, J., and Odem, W. (2000). Membranes for the control of
natural organic matter from surface waters. Water Research, 34(13), p. 3355-3370.
Scott, B. F., Mactavish, D., Spencer, C., Strachan, W. M., and Muir, D. C. (2000).
Haloacetic acids in Canadian lake waters and precipitation. Environmental Science and
Technology, 34(20), p. 4266-4272.
Scott, B. F., and Alaee, M. (1998). Determination of haloacetic acids from aqueous
samples collected from the Canadian environment using an in situ derivatization
technique. Water Quality Research Journal of Canada, 33(2), p. 279-293.
180
Seidel, C., Frey, M. M., and Summer, R. S. (2000). Understanding disinfection byproduct behaviour in water distribution systems. In: Proceedings of the American Water
Works Association Water Quality Technology Conference, November 7-12, Denver,
CO, US.
Sérodes, J.-B., Rodriguez, M. J., Li, H., and Bouchard, C. (2003). Occurrence of THMs
and HAAs in experimental chlorinated waters of the Quebec City area (Canada).
Chemosphere, 51(4), p. 253-263.
Sharp, E. L., Parson, S. A., and Jefferson, B. (2006). Coagulation of NOM: linking
character to treatment. Water Science and Technology, 53(7), p. 67-76.
Sharp, E. L., Parsons, S. A., and Jefferson, B. (2006a). The impact of seasonal
variations in DOC arising from a moorland peat catchment on coagulation with iron and
aluminium salts. Environmental Pollution, 140(3), p. 436-443.
Singer, P. C. (2006). Regulation of only five haloacetic acids is neither sound science
nor good policy. In: Proceedings of the Symposium on Safe Drinking Water: Where
Science Meets Policy, March 16-17, Chapel Hill, NC, US.
Singer, P. C., Weinberg, H. S., Brophy, K., Liang, L., Roberts, M., Grisstede, I.,
Krasner, S. W., Baribeau, H., Arora, H., and Najm, I. (2002). Relative dominance of
HAAs and THMs in treated drinking water, American Water Works Association
Research Foundation report no. 90844, Denver, CO, US.
Singer, P. C. (1999). Formation and Control of Disinfection By-Products in Drinking
Water, edited by Singer, P.C., American Water Works Association, Denver, CO, US.
Singer, P. C., Obolensky, A., and Greiner, A. (1995). DBPs in chlorinated North
Carolina drinking waters. Journal of the American Water Works Association, 87(10), p.
83-92.
Sinha, S., Sohn, J., and Amy, G. (1997). Reactivity of NOM in forming chlorinated
DBPs. In: Proceedings of the American Water Works Association Annual Conference,
June 15-19, Atlanta, GA, US.
Skelly, N. E. (1982). Separation of inorganic and organic anions on reversed-phase
liquid chromatography columns. Analytical Chemistry, 54(4), p. 712-715.
Sohn J., Amy G., Cho J., Lee Y., and Yoon Y. (2004). Disinfectant decay and
disinfection by-products formation model development: Chlorination and ozonation byproducts, Water Research, 38(10), p. 2461-2478.
Speitel G. E. (1999). Control of disinfection by-product formation using chloramines
Formation and control of disinfection by-products in drinking water, edited by Singer,
P.C., American Water Works Association, Denver, CO, US.
Sung, W., Reilley-Matthews, B., O’Day, D. K., and Horrigan, K. (2000). Modeling
DBP formation. Journal of the American Water Works Association, 92(5), p. 53-63.
181
Symons, J. M. (1999). Disinfection by-products: A historical perspective. Formation
and control of disinfection by-products in drinking water, edited by Singer, P.C.,
American Water Works Association, Denver, CO, US.
Symons, J. M., Krasner, S. W., Simms, L. A., and Sclimenti, M. (1993). Measurement
of THM and precursor concentrations revisited: The effect of bromide ion. Journal of
the American Water Works Association, 85(1), p. 51-62.
Trussell, R. R., Umphres, M. D. (1978). Formation of Trihalomethanes. Journal of the
American Water Works Association, 70(11), p. 604-612.
Tung, H.-H., Unz, R. F., and Xie, Y. F. (2006). HAA removal by GAC adsorption.
Journal of the American Water Works Association, 98(6), p. 107-112.
Ueno, H., Moto, T., Sayato, Y., and Nakamuro, K. (1996). Disinfection by-products in
the chlorination of organic nitrogen compounds: By-products from kynurenine.
Chemosphere, 33(8), p. 1425-1433.
Urano, K., and Takemasa, T. (1986). Formation equation of halogenated organic
compounds when water is chlorinated. Water research, 20(12), p. 1555-1560.
Urbansky, E. T. (2000). Techniques and methods for the determination of haloacetic
acids in potable water. Journal of Environmental Monitoring, 2(4), p. 285-291.
US EPA (2003). Method 552.3: Determination of haloacetic acids and dalapon in
drinking water by liquid-liquid microextraction, derivatization and gas chromatography
with electron capture detection, Revision 1.0, Technical Support Centre, Office of
Ground Water and Drinking Water, Cincinnati, OH, US.
US EPA (1998). National Primary Drinking Water Regulations: Disinfectants and
Disinfection Byproducts, Final Rule, Federal Register, 63:241:69390.
drinking water by liquid-liquid extraction, derivatization and gas chromatography with
electron capture detection, Revision 1.0, National Exposure Research Laboratory,
Office of Research and Development, Cincinnati, OH, US.
US EPA (1995a). Method 551.1: Determination of chlorination disinfection byproducts,
chlorinated solvents, and halogenated pesticides/herbicides in drinking water by liquidliquid extraction and gas chromatography with electron-capture detection, Revision 1.0,
National Exposure Research Laboratory, Office of Research and Development,
Cincinnati, OH, US.
US EPA (1995b). Method 502.2: Volatile organic compounds in water by purge and
trap capillary column gas chromatography with photoionization and electrolytic
conductivity detectors in series, Revision 2.1, National Exposure Research Laboratory,
182
US EPA (1995c). Method 524.2: Measurement of purgeable organic compounds in
water by capillary column gas chromatography/mass spectrometry , Revision 4.1,
National Exposure Research Laboratory, Office of Research and Development,
Cincinnati, OH, US.
drinking water by ion-exchange liquid-solid extraction and gas chromatography with an
electron capture detector, Revision 1.0, Environmental Monitoring Systems Laboratory,
US EPA (1979). National Interim Primary Drinking Water Regulations: Control of
Trihalomethanes in Drinking Water, Federal Register. 44:231:68624.
Valentine, R. L. (1986). Bromochloramine oxidation of N,N-diethyl-pphenylenediamine in the presence of monochloramine. Environmental Science and
Technology, 20(2), p. 166-170.
Vichot, R., and Furton, K. G. (1994). Variables influencing the direct determination of
haloacetic acids in water by reversed-phase ion-pair chromatography with indirect UV
detection. Journal of Liquid Chromatography, 17(20), p. 4405-4429.
Vikesland, P. J., Ozekin, K., and Valentine, R. L. (2001). Monochloramine decay in
model and distribution system waters. Water Research, 35(7), p. 1766-1776.
Vikesland, P. J., Ozekin, K., and Valentine, R. L. (1998). Effect of natural organic
matter on monochloramine decomposition: Pathway elucidation through the use of mass
and redox balances. Environmental Science and Technology, 32(10), p. 1409-1416.
Vilgé-Ritter, A., Maison, A., Boulangé, T., Rybacki, D., and Bottero, J.-Y. (1999).
Removal of natural organic matter by coagulation-flocculation: A pyrolysis-GC-MS
study. Environmental Science and Technology, 33(17), p. 3027-3032.
Villanueva, C. M, Kogevinas, M., and Grimalt, J. O. (2003). Haloacetic acids and
trihalomethanes in finished drinking waters from heterogeneous sources. Water
Research, 37(4), p. 953-958.
Volk, C., Bell, K., Ibrahim, E., Verges, D., Amy, G., and Lechevallier, M. (2000).
Impact of enhanced and optimised coagulation on removal of organic matter and its
biodegradable fraction in drinking water. Water Research, 34(12), p. 3247-3257.
Von Gunten, U., and Oliveras, Y. (1998). Advanced oxidation of bromide-containing
waters: Bromate formation mechanisms. Environmental Science and Technology, 32(1),
p. 63-70.
Waller, K., Swan, S. H., DeLorenze, G., and Hopkins, B. (1998). Trihalomethanes in
drinking water and spontaneous abortion. Epidemiology, 9(2), p. 134-140.
Wang, W., Ye, B., Yang, L., Li, Y., and Wang, Y. (2007). Risk assessment on
disinfection by-products of drinking water of different water sources and disinfection
processes. Environment international, 33(2), p. 219-225.
183
Watson, W. (1993). Mathematical modelling of the formation of THMs and HAAs in
chlorinated natural waters, American Water Works Association Research Foundation
report, Denver, CO, US.
Weinberg, H. S., Krasner, S. W., Richardson, S. D., and Thruston, A. D., Jr. (2002). The
Occurrence of Disinfection By-Products (DBPs) of Health Concern in Drinking Water:
Results of a Nationwide DBP Occurrence Study, EPA/600/R-02/068, National Exposure
Research Laboratory, Office of Research and Development, Athens, GA, US.
Weinberg, H. S., Glaze, W. H., Krasner, S. W., and Sclimenti, M. J. (1993). Formation
and removal of aldehydes in plants that use ozonation. Journal of the American Water
Westerhoff, P., Chao, P. and Mash, H. (2004). Reactivity of natural organic matter with
aqueous chlorine and bromine. Water Research, 38(6), p. 1502-1513.
Westerhoff, P., Debroux, J., Amy, G. L., Gatel, D., Mary, V., and Cavard, J. (2000).
Applying DBP models to full-stale plants. Journal of the American Water Works
Association, 92(3), p. 89-102.
White, D. M., Garland, D. S., Narr, J., and Woolard, C. R. (2003). Natural organic
matter and DBP formation potential in Alaskan water supplies. Water Research, 37(4),
p. 939-947.
White, G.C. (1999). Handbook of Chlorination and Alternative Disinfectants, 4th
Edition, edited by John Wiley and Sons, Inc., New York, NY, US.
World Health Organisation (WHO) (2006). Guidelines for drinking-water quality, First
addendum to third edition, vol. 1, Recommendations.
World Health Organisation (WHO) (2000). Disinfectants and Disinfection By-Products,
available at http://whqlibdoc.who.int/hq/2000/a68673_guidelines_2.pdf.
World Health Organisation (WHO) (1998). Chlorine, International Programme on
Chemical
Safety
Poisons,
available
at
http://www.inchem.org/documents/pims/chemical/pim947.htm.
Williams, D. T., LeBel, G. L., and Benoit, F. M. (1997). Disinfection by-products in
Canadian drinking water. Chemosphere, 34(2), p. 299-316.
Winterton, N. (2000). Chlorine: The only green element - Towards a wider acceptance
of its role in natural cycles. Green Chemistry, 2(5), p. 173-225.
Wolfe, R.L., Ward, N. R., and Olson, B. H. (1984). Inorganic chloramines as drinking
water disinfectants: a review. Journal of the American Water Works Association,
76(55), p.74–88.
184
Woo, Y. T., Lai, D., and McLain, J. L. (2002). Use of mechanism-based structureactivity relationships analysis in carcinogenic potential ranking for drinking water
disinfection by-products. Environmental Health Perspectives, 110(S1), p. 75-87.
Xie, Y. (2003). Disinfection by-products in drinking water: Formation, analysis and
control, edited by Lewis Publishers, Chelsea, MI, US.
Xie, Y., Rashid, I., Zhou, H. J., and Gammie, L. (2002). Acidic methanol methylation
for HAA analysis: Limitations and possible solutions. Journal of the American Water
Xie, Y. (2001). Analyzing haloacetic acids using gas chromatography/mass
spectrometry. Water research, 35(6), p. 1599-1602.
Xie, Y., Zhou, H., and Romano, J. P. (2000). Development of a capillary electrophoresis
method for haloacetic acids. In: Natural Organic Matter and Disinfection By-Products:
Characterization and Control in Drinking Water. American Chemical Society
Symposium Series 761, edited by Sylvia E. Barrett, Stuart W. Krasner and Gary L.
Amy. Washington, DC, US, p. 356-365.
Xie, Y., and Romano, J. P. (1997). Application of capillary ion electrophoresis for
drinking water analysis. In: Proceedings of the American Water Works Association
Water Quality Technology Conference, November 7-12, Denver, CO, US.
Yang, X., Shang, C., and Westerhoff, P. (2007). Factors affecting formation of
haloacetonitriles, haloketones, chloropicrin and cyanogen halides during
chloramination. Water Research, 41(6), p. 1193-1200.
Young, M. S., Mauro, D. M., Uden, P. C., and Reckhow, D. A. (1995). The formation
of nitriles and related halogenated disinfection by-products in chlorinated and
chloraminated water; applications of microscale analytical procedures. American
Chemical Society (Environmental Chemistry Division Preprints), 35(2), p. 748-751.
Zhang, X., Echigo, S., Minear, R. A., and Plewa, M. J. (2000). Characterization and
comparison of disinfection by-products of four major disinfectants. In: Natural Organic
Matter and Disinfection By-Products: Characterization and Control in Drinking Water.
American Chemical Society Symposium Series 761, edited by Sylvia E. Barrett, Stuart
W. Krasner and Gary L. Amy. Washington, DC, US, p. 299-314.
Zhuo, C., Chengyong, Y., Junhe, L., Huixian, Z., and Jinqi, Z. (2001). Factors on the
formation of disinfection by-products MX, DCA and TCA by chlorination of fulvic acid
from lake sediments. Chemosphere, 45(3), p. 379-385.
185

Cynthia BOUGEARD Haloacetic Acids and other Disinfection By

Transcription

Similar documents

Typical education package CNC machines. Haas CNC the choice

C apabilities

SOLIKLOOR

Edition 6 - Malkara School

ST-20Y Datasheet - Haas Automation®, Inc.

waterfront dream home in waters edge granbury texas

TL-2 Datasheet - Haas Automation®, Inc.

Chlorine Dioxide monitor - Chemical Support Systems

ppt - The 2016 ACM SIGMOD/PODS Conference

Chlorine-Free Water Throughout Your Home